Laser Chemistry of Organometallics 9780841226876, 9780841213869, 0-8412-2687-3

Citation preview

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.fw001

Laser Chemistry of Organometallics

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.fw001

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

ACS SYMPOSIUM SERIES

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.fw001

Laser Chemistry of Organometallics J. Chaiken, EDITOR Syracuse University

Developed from a symposium sponsored by the Division of Physical Chemistry and the Division of Inorganic Chemistry, Inc., at the 203rd National Meeting of the American Chemical Society, San Francisco, California, April 5-10, 1992

American Chemical Society, Washington, DC 1993

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

530

Library of Congress Cataloging-in-Publication Data Laser chemistry of organometallics: developed from a symposium sponsored by the Division of Physical Chemistry and the Division of Inorganic Chemistry, Inc., at the 203rd National Meeting of the American Chemical Society, San Francisco, California, April 5-10, 1992; J. Chaiken, editor. p.

cm.—(ACS symposium series, ISSN 0097-6156; 530)

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.fw001

Includes bibliographical references and index. ISBN 0-8412-2687-3 1. Laser photochemistry—CHEMISTRY—Congresses. 2. Organometallic compounds—Congresses. I. Chaiken, J., 1955— . II. American Chemical Society. Division of Physical Chemistry. III. American Chemical Society. Division of Inorganic Chemistry. IV. American Chemical Society. Meeting (203rd: 1992: San Francisco, Calif.) V. Series. QD716.L37L37 1993 547.050455—dc20

93-7877 CIP

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences—Permanence of Paper for Printed Library Materials, ANSI Z39.48-1984. @ Copyright © 1993 American Chemical Society A l l Rights Reserved. The appearance of the code at the bottom of the first page of each chapter in this volume indicates the copyright owner's consent that reprographic copies of the chapter may be made for personal or internal use or for the personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay the stated per-copy fee through the Copyright Clearance Center, Inc., 27 Congress Street, Salem, M A 01970, for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to copying or transmission by any means—graphic or electronic—for any other purpose, such as for general distribution, for advertising or promotional purposes, for creating a new collective work, for resale, or for information storage and retrieval systems. The copying fee for each chapter is indicated in the code at the bottom of the first page of the chapter. The citation of trade names and/or names of manufacturers in this publication is not to be construed as an endorsement or as approval by ACS of the commercial products or services referenced herein; nor should the mere reference herein to any drawing, specification, chemical process, or other data be regarded as a license or as a conveyance of any right or permission to the holder, reader, or any other person or corporation, to manufacture, reproduce, use, or sell any patented invention or copyrighted work that may in any way be related thereto. Registered names, trademarks, etc., used in this publication, even without specific indication thereof, are not to be considered unprotected by law. PRINTED IN T H E UNITED STATES O F A M

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

1993 Advisory Board ACS Symposium Series M . Joan Comstock, Series Editor V . D e a n Adams University of Nevada— Reno

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.fw001

Robert J. A l a i m o Procter & Gamble Pharmaceuticals, Inc. Mark Arnold University of Iowa D a v i d Baker University of Tennessee A r i n d a m Bose Pfizer Central Research Robert F . Brady, Jr. Naval Research Laboratory Margaret A. Cavanaugh National Science Foundation Dennis W. Hess Lehigh University

Bonnie Lawlor Institute for Scientific Information Douglas R . L l o y d The University of Texas at Austin Robert M c G o r r i n Kraft General Foods Julius J. M e n n Plant Sciences Institute, U.S. Department of Agriculture Vincent Pecoraro University of Michigan Marshall Phillips Delmont Laboratories George W . Roberts North Carolina State University A . Truman Schwartz Macalaster College

H i r o s h i Ito IBM Almaden Research Center

John R . Shapley University of Illinois at Urbana-Champaign

Madeleine M. Joullie University of Pennsylvania

L . Somasundaram Ε. I. du Pont de Nemours and Company

Gretchen S. K o h l Dow-Corning Corporation

Peter Willett University of Sheffield (England)

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

Foreword

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.fw001

THE ACS S Y M P O S I U M S E R I E S was first published in 1974 to

provide a mechanism for publishing symposia quickly i n book form. The purpose of this series is to publish comprehensive books developed from symposia, which are usually "snapshots i n time" of the current research being done o n a topic, plus some review material o n the topic. F o r this reason, it is necessary that the papers be published as quickly as possible. Before a symposium-based book is put under contract, the proposed table of contents is reviewed for appropriateness to the topic and for comprehensiveness of the collection. Some papers are excluded at this point, and others are added to r o u n d out the scope of the volume. I n addition, a draft of each paper is peer-reviewed prior to final acceptance or rejection. This anonymous review process is supervised by the organizer(s) of the symposium, who become the editor(s) of the book. T h e authors then revise their papers according to the recommendations of both the reviewers and the editors, prepare camera-ready copy, and submit the final papers to the editors, who check that all necessary revisions have been made. A s a rule, only original research papers and original review papers are included i n the volumes. Verbatim reproductions of previously published papers are not accepted. M. Joan Comstock Series E d i t o r

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.pr001

Preface S C I E N T I F I C A L L Y , E C O N O M I C A L L Y , A N D P O L I T I C A L L Y , this book and the symposium on which it is based mark a turning point that's been a long time in coming. Basic and applied research in laser chemistry is being directed to a broader range of possible advantages of using lasers as agents of chemical and physical change. The initial surge of interest in lasers as "molecular scalpels" began in the 1960s and now has evolved into a list of types of laser chemistry. One result of this evolution is the perception that metal-containing systems, and in particular organometallic systems, have characteristics that are sometimes uniquely synergistic with the basic properties of laser excitation. This book appraises the current state of the laser chemistry of organometallic systems within the conceptual framework presented in the first chapter. Fundamental research is best approached by choosing to obtain the most detailed information possible on the simplest possible chemical and physical systems with the greatest control of all relevant parameters. This ideal may be considered a luxury in the future because applied research often requires making the best judgments possible based on the limited information at hand. Fundamental research may become more difficult to pursue in the future and certainly not because all important questions have been answered or even addressed. This situation is a simple consequence of the economic and political pressure to turn our attention to the study of chemical and physical systems with near-term commercial value. For the past few decades, the U.S. government has fostered the growing perception that for economic and political purposes, it is time to harvest our cache of fundamental knowledge. Interest in the laser chemistry of organometallic systems is not only the result of the desire for new commercial applications; it is also a natural consequence of the desire to observe new phenomena and thereby understand the applicability of our current fundamental pictures to more complex phenomena. That is, simple curiosity drives this research also. Systems that involve large, complicated molecules in dispersed and condensed media are being studied using methods previously reserved for the study of atoms and diatomics under collisionless conditions. Conceptually, the importance of studying the transition from fundamentally simple to fundamentally complex systems, for purposes of deducing underlying scaling laws and the means to manipulate matter and energy in new ways, is important to understanding why the research in this field is beginning to escalate, why the research in this field is important, and why there is a need for the information to be presented now. This book attempts to survey a broad range of chemical and physical research, fundamental and applied, that defines the current state of laser chemistry of organometallic systems. The book's scope is intentionally expansive to provide an introduction to laser chemistry for scientists and engineers with the

xi

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.pr001

widest possible range of interests. One of my intentions was to provide an entry point into the primary literature of laser chemistry that spans the catalysis, ceramics, microelectronics, photonics, and materials research fields. It should be useful to undergraduate and graduate students who want an introduction to the field of laser chemistry as well as administrators and managers who want an overview into the field as it currently stands. The level of presentation should be accessible to readers with an advanced undergraduate background in physical or inorganic chemistry. This book should be unusual and useful to many specialists in laser chemistry in the sense that I have invited the authors to speculate more than is customary in the primary literature. I have invited them to disagree, and I have made no effort to obtain a consensus on anything because the field is still rapidly evolving. I asked only that they make an effort to identify some connection with the conceptual framework I present in the first chapter. Some presentations in the symposium were modified because of confidentiality requirements of the scientists' industrial employers. For these chapters, my judgment was in the spirit of applied research: It's better to have some information than none at all.

Acknowledgments Many companies and organizations made generous financial contributions. Without them, the symposium would not have been possible: Coherent, Lambda Physik, Lumonics, E G & G - P A R , the Petroleum Research Fund as administered by the American Chemical Society, the Divisions of Inorganic Chemistry, Inc., and Physical Chemistry of the American Chemical Society, and Syracuse University's Vice President for Research, Ben Ware. Between the time the symposium was planned and the time it occurred, the Russian ruble was devalued three times and the Canadian government enacted a spending freeze. The generosity of the sponsors literally saved 25% of the symposium. In strategies and tactics, Eric Weitz and J. J. Valentini made many helpful suggestions throughout. I am particularly grateful to the authors and participants in the symposium who have done an excellent job in educating me and each other. Because of a severe family emergency that hampered my editing, I appreciate the understanding of all those at A C S Books who were involved in the preparation of this book. J. CHAIKEN

Syracuse University Syracuse, NY 13244-4100 December 1, 1992

xii

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

Chapter 1

Laser Chemistry of Organometallic Species Conceptual Framework and Overview J. Chaiken

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch001

Department of Chemistry, Syracuse University, Syracuse, NY 13244-4100

"Laser Chemistry" refers to chemical processes initiated by the action of laser(s) on matter. Because of the rapid advancement of relevant fundamental and applied research over the last several years, a conference to assess the state of the field is appropriate at this time. One emerging realization is that a unique synergism exists between the fundamental properties of laser excitation and the photochemistry of organometallics so Laser Chemistry of Organometallics was chosen as the specific focus of this Symposium. In succeeding chapters, accounts of research in multiphoton processes, metal atom chemistry, the chemistry of coordinatively unsaturated organometallics, thin film deposition, and the formation and chemistry of ultrafine particles and clusters in gas and condensed phases will all be presented within the unifying concept of Laser Chemistry of Organometallics. Our goal in this chapter is to describe this conceptual framework. The properties of lasers and laser light allow reaction mixtures to be energized selectively with respect to species, quantum state(s), and temporal and/or spatial distribution. The composition and physical properties of organometallic chemical systems can lead to unique phenomena and processes when lasers are used to energize systems. In the context of specific forms of laser chemistry, we shall discuss some particular examples in which organometallics provide the possibilities for unique chemistry. The following chapters describe research which bears on the processes and phenomena connected in the flowchart in Figure 1. Some of the fundamental and practical questions which must be answered to discover the extent of possible applications are evident. Introduction to Laser Chemistry The term "Laser Chemistry" is nearly as old as the laser itself with reviews(1,2,3,4) and patents(5,6) first appearing in the literature in the sixties(7) and continuing(8) today. Of the good recent reviews available, we choose to mostly follow the

0097-6156/93/0530-0001$07.00/0 © 1993 American Chemical Society In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch001

2

LASER CHEMISTRY O F ORGANOMETALLICS

categories and terminology used by Woodrufff 1). To survey the unique possibilities of laser chemistry as applied to organometallics, we first present a very short introduction to laser chemistry in general, then a short digression into the established single(9) and multiphoton(70)chemistry of organometallics, and finally a brief summary of some very recent results of our attempts to synthesize clusters and ultrafine particles using laser chemistry of organometallics. For clarity in presentation, we mostly choose specific examples from our own research to illustrate our main points. No attempt is made to be exhaustive of the literature. The flowchart on the next page is meant to suggest a certain temporal separation in a sequence of events, which taken together, could be called "laser chemistry of organometallics." The first stage is the laser excitation process. The second stage is the chemistry of the nascent distribution of species and states. A final stage describes still longer time scale chemistry and is convenient if one is concerned about producing materials which are subsequently removed from a reaction vessel and subjected to some other processing or aging in the course of characterization. This could also correspond to sintering in the case of ceramic particle chemistry or possibly to post-deposition annealing in air in the case of transparent metal film deposition. Further discussion and examples of each stage follow. The chart is labeled to suggest that the process can occur either homogeneously in gas or condensed phases, or heterogeneously involving a number of phases. These questions of phase set the context of the mass and energy transfer characteristics for the laser chemistry. Thus terms like "film-phase" arise in the application of the flowchart to processes like laser chemical vapor deposition. Although not necessarily involving an organometallic, the same scenario can be applied to the use of laser ablation to rapidly produce gas phase metal atoms from a metal rod(see Hackett's chapter). Although the laser provides a unique form of excitation, much of the chemistry which occurs after laser induced species begin to exist is dictated by kinetics and thermodynamics consistent with those mass(7i) and energy transfer(72) conditions. Bauerle's chapter give an introduction and analysis of some of these issues with regard to modelling such systems. Singmaster's chapter presents the current state of the well studied process of laser chemical vapor deposition of metal films using metal carbonyl precursors. The net result of the laser excitation is to convert an organometallic into either metal atoms, metal ions, or coordinatively unsaturated species, or some mixture of all of the above and ligands, which can then participate in materials chemistry or catalysis. Questions associated with mechanisms for unimolecular chemistry(73,74,75,76,77,) are therefore of paramount importance during the first stage because they determine the internal state distributions of the atoms which then may participate in bimolecular chemistry. If ionic species are to be used as reactants, they could be generated directly via a uv-visible multiphoton process. Laser induced d i s c h a r g e s ' ^ ) or other methodsf20), are also available in which the ions are mostly formed by electron impact phenomena. The internal state and species distribution of these nascent ionic species would often be interpreted in terms of quasi-equilibrium theory(QET). In any case, some recently explored spectroscopy of organometallic ions is described in Kimura's chapter. In the single to few photon range, IR excitation can only lead to the kinds of laser chemistry Woodruff calls either selective heating or vibrationally enhanced. Thermodynamics dictates that IR laser excitation can only lead to dissociation via nearly simultaneous absorption of at least a few photons and such multiphoton dissociation tends to produce neutral fragments. Visible and shorter wavelength excitation tends to produce ions in the multiphoton limit. In the single photon limit,

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

Conceptual Framework and Overview

CHAIKEN

Fragmentation of gas/film phase organometallic species Thermolysis

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch001

Photolysis

RRKM, IVR, IER, QET, internal energy redistribution 'direct' dissociation excited state production incomplete energy redistribution large translational energy release

ft 1

K* CO transition, the CT transition. The two excitation cases are discussed separately below and the schematic energy levels are shown in Figure 4. For the CT band experiments at least six photons are required to supply sufficient energy to both dissociate Fe from the complex and ionize the atom. Measured power indices indicated that one or more steps in the process were saturated. Relative populations of the photoproduced Fe electronic states were then compared as a function of organic ligand to discern the dynamical constraints on the photosystem. The most intense REMPI spectral lines were observed between 447.2 and 448.4 nm and quantitative efforts were concentrated in this region of the spectrum. The values of the integrated areas of the lines belonging to the e Dj 90 kcal/mol, while the C-N bond is only 72 kcal/mol) (32). The experimental procedure

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

8. XU ET AL.

Laser Chemistry Relevant to III—V Semiconductor Growth

was simply to photolyze C 2 H 5 N D 2 , for example, and look for H versus D atom formation. At lower laser powers (< several MW/cm ), virtually all atomic hydrogen results from N-D bond cleavage (30). With sufficient laser power, one can see the CH channel "turn on." Power dependence studies indicate that a two-photon process is responsible for H-atom production, while N-D bond cleavage is produced through a one-photon route. In the sense of absolute H-atom production referenced to N H 3 (where the quantum yield is -1.0), H-atom generation from C 2 H 5 N H 2 is -30-50% that resulting from NH3. (Note that differences in absorption cross sections have been taken into account.) Consequently, one can conclude that N - H bond cleavage is fairly competitive with C-N bond cleavage, despite the difference in bond strength (32). Results on the 193 nm photolysis of C2H5ASH2 are shown in Figure 3, and it is clear from the broadening of the H-atom Doppler profile that a second H-atom channel is "turning on" dramatically at higher laser powers (31). Since the H-atom profile is broadening significantly, it is likely that the important intermediate photolysis step involves irradiation of AsH2. Studies are currendy being planned that will involve the selectively-deuterated compounds C2D5ASH2 and C2H5ASD2. In this case, positive identification of the reactive site is possible. It must also be noted that the absolute amount of H-atom production resulting from M E A s photolysis is comparable to that for 193 NH3 photolysis, where the H-atom quantum yield is -1.0. Consequendy, the prospects for inducing macroscopic H-atom production in M E A s growth environments are encouraging.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch008

2

Enhancing H-atom Production. To influence the overall photochemistry of C2H5NH2, we propose to use two photolysis lasers operating at 193 and 248 nm. By using this additional 248 nm photon source, one can alter the 193-nm induced H/D ratio that results from C2H5ND2 photolysis (33). As shown in our recent two-color experiment on CICH2CH3, H-atom production from the ethyl radical can be enhanced by using 248 nm radiation (27). By using this second photolysis source operating at 248 nm, we can also photolytically increase H-atom production from a C2H5 radical produced via the 193 nm photolysis of C2H5ND2 (33).

Significant H-atom enhancement is also possible with TEAs and M E A s (33). Figure 4a shows H-atom Doppler profiles at Lyman-oc produced via TEAs photolysis using three different photolysis combinations. Clearly, H-atom formation arising from a combination of 193 and 248 nm laser pulses is enhanced significantly when compared with H-atom generation resulting from the sum of the two photolysis lasers used separately. Since the absorption cross sections for TEAs at 193 nm and 248 nm are reasonable (1.76 x 1 0 ' c m and 1.2 x 1 0 ' c m , respectively) (34), it is not surprising that H-atom formation is observed in each individual case. However, the net H-atom signal is enhanced by a factor of 5 when compared with the H-atom signal resulting from the sum of the two lasers acting alone. We submit that in the case of TEAs, the desired reaction (C2H5 —> C2H4 + H) is being photolytically enhanced by the 248 nm laser pulse. In contrast to TEAs, MEAs has direct H-atom formation channels available. Nonetheless, we still expect that 193 nm excitation of M E A s will induce significant As-C dissociation in addition to A s - H bond cleavage, since the A s - C bond is significantly weaker than the As-H bond (35). If As-C bond cleavage does in fact occur, then ethyl radicals will be produced. Will 248 nm radiation enhance H-atom formation following 193 nm M E A s photolysis in a fashion similar to the case of 193 nm TEAs photolysis? The results of such an experiment can be found in Figure 4b. When compared with the TEAs experiment, the main difference for M E A s is the fact 17

2

18

2

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

111

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch008

112

LASER CHEMISTRY OF ORGANOMETALLICS

-7

Vo +7 VUV Probe Frequency (cm- ) 1

Figure 3. H-atom Doppler profiles resulting from the 193 nm photolysis of C 2 H 5 A S H 2 at two photolysis powers. The larger photolysis power has clearly activated a second H-atom channel, as evidenced by the Doppler broadening. Here, v = 82,259.1 cm" . 1

0

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

8.

X U E T AL.

113 Laser Chemistry Relevant to HI—V Semiconductor Growth

•

193.248

•

248 alone

O

193 alone

S

#» ^

TEAs

£

a)

•

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch008

o

- 6 - 3

V

3

0

6 1

V U V Probe Frequency (cm" )

Figure 4a. H-atom Doppler profiles resulting from (C2H5)3As photolysis using three different excitation schemes: 193 nm alone, 248 nm alone, and 193 nm followed by 248 nm. It is clear that significant enhancement of the H-atom signal occurs when both lasers are used. Here, VQ is equal to Lyman-oc (82,259 cm"*).

O

193 alone

•

O

- 6

- 3

b)

O

Vo

3

6 -1

V U V Probe Frequency (cm )

Figure 4b. H-atom Doppler profiles resulting from C2H5ASH2 photolysis using two different excitation schemes: 193 nm alone and 193 nm followed by 248 nm. (When only 248 nm radiation was used, there was no discernible H-atom signal.) It is clear that some enhancement of the H-atom signal occurs when both lasers are used. Here, V Q is equal to Lyman-a (82,259 cm *). -

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

114

LASER CHEMISTRY O F ORGANOMETALLICS

that 248 nm radiation alone produced no discernible H-atom signal, probably due to a low absorption cross section at this wavelength. However, the H-atom signal is enhanced (albeit slightly) when 193 nm and 248 nm pulses are used together as opposed to 193 nm radiation acting alone. The actual enhancement is by -50%. If this enhancement is occurring via the ethyl radical, then we are achieving the desired photochemistry. Once again, experiments with selectively-deuterated M E A s compounds should be particularly useful for addressing this issue. Finally, we reemphasize that understanding the photochemistry of group V ethyl-containing compounds may ultimately prove important in understanding and/or enhancing n i - V semiconductor growth processes that utilize metalorganic compounds. In this area, MEAs is a particularly attractive candidate since it contains only one ethyl group.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch008

Acknowledgments Acknowledgment is made to the Donors of the Petroleum Research Fund, administered by the American Chemical Society, for partial support of this research. We also acknowledge support by the Louisiana Board of Regents, the Center for Bioenvironmental Research at Tulane University, and the Department of Energy via its NIGEC program. Special thanks go to Shannon Koplitz for presentation assistance.

Literature Cited 1. Manasevit, H.M. J. Cryst. Growth 1981, 55, 1. 2. Dapkus P.D.; Manasevit H.M.; Hess K.L.; Low T.S.; Stillman G.E.J.Cryst. Growth 1981, 55, 10. 3. Tsang W.T. Appl. Phys. Lett. 1984, 45, 1234. 4. Tsang W.T.; Chiu T.H.; Cunningham J.E.; Robertson A. Appl. Phys. Lett. 1987, 50, 1376. 5. Tsang W.T.; Chiu T.H.; Cunningham J.E.; Robertson A. J. Appl. Phys. 1987, 62, 2302. 6. Robertson A.; Chiu T.H.; Tsang W.T.; Cunningham J.E. J. Appl. Phys. 1988, 64, 87. 7. Putz N.; Heinecke H.; Heyen M.; Balk P.; Weyers M.; Luth H. J. Cryst. Growth 1986, 74, 292. 8. Kukimoto H.; Ban Y.; Komatsu H.; Takechi M.; Ishizaki M. J. Cryst. Growth 1986, 77, 223. 9. Nishizawa J.; Kurabayashi T.; Hoshina J. J. Electrochem. Soc. 1987, 502. 10. McCaulley J.A.; McCrary V.R.; Donnelly V.M. J. Phys. Chem. 1989, 93, 1148. 11. Donnelly V.M.; McCrary V.R.; Appelbaum A.; Brasen D.; Lowe W.P. J. Appl. Phys. 1987, 61, 1410. 12. Tu C.W.; Donnelly V.M.; Beggy J.C.; Baiocchi F.A.; McCrary V.R.; Harris T.D.; Lamont M.G. Appl. Phys. Lett. 1988, 52, 966. 13. Donnelly V.M.; Brasen D.; Appelbaum A.; Geva M. J. Appl. Phys. 1985, 58, 2022. 14. Nishizawa J.; Kurabayashi T.; Abe H.; Sakurai N. J. Vac. Sci. Technol. A 1987, 5, 1572. 15. Chen C.H.; Larsen C.A.; Stringfellow G.B. Appl. Phys. Lett. 1987, 50, 219. 16. Lum R.M.; Klingert J.K.; Lamont M.G. Appl. Phys. Lett. 1987, 50, 284. 17. Lum R.M.; Klingert J.K.; Wynn A.S.; Lamont M.G. Appl. Phys. Lett. 1988, 52, 1475. 18. Hummel S.G.; Zou Y.; Beyler C.A.; Grodzinski P.; Dapkus P.D.; McManus

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch008

8. XU ET AL.

Laser Chemistry Relevant to III-V Semiconductor Growth

115

J.V.; Zhang Y.; Skromme B.J.; LeeW.I.Appl. Phys. Lett. 1992, 60, 1483. 19. Speckman D.M.; Wendt J.P. Appl. Phys. Lett. 1990, 56, 1134. 20. Zacharias H.; Rottke H.; Danon J.; Welge K.H. Opt. Commun. 1981, 37, 15. 21. Brum J.L.; Deshmukh S.; Koplitz B. J. Chem. Phys. 1991, 95, 2200. 22. Xu X.; Deshmukh S.; Brum J.L.; Koplitz B. Appl. Phys. Lett. 1991, 58, 2309. 23. DenBaars S.P.; Maa B.Y.; Dapkus P.D.; Melas A. J. Electrochem. Soc. 1990, 738,2068. 24. Brum J.L.; Deshmukh S.; Koplitz B. J. Chem. Phys. 1990, 93, 7504. 25. Deshmukh S.; Brum J.L.; Koplitz B. Chem. Phys. Lett. 1991,176,198. 26. Arthur N. J. Chem. Soc. Faraday. Trans. 1986, 82, 1057. 27. Robin M.B. Higher Excited States of Polyatomic Molecules; Academic Pres New York, NY, 1974; Vol. 1. 28. Kawasaki M.; Kasatani K.; Sato H.; Shinohara H.; Nishi N. Chem. Phys. 1984, 88, 135. 29. Brum J.L.; Deshmukh S.; Koplitz B. J. Phys. Chem. 1991, 95, 8676. 30. Xu X.; Deshmukh S.; Brum J.L.; Koplitz B. Chem. Phys. Lett. 1992, 188, 32. 31. Xu X.; Brum J.L.; Deshmukh S.; Koplitz B. J. Phys. Chem. 1992, 96, 2924. 32. Sanderson R.T. Chemical Bonds and Bond Energy; Academic Press: New York, NY, 1971, 153. 33. Xu X.; Wang Z.; Brum J.L.; Yen Y.-F.; Deshmukh S.; Koplitz B.J.Phys. Chem. 1992, 96, 7048. 34. McCrary V.R.; Donnelly V.M. J. Cryst. Growth 1987, 84, 253. 35. Berkowitz J. J. Chem. Phys. 1988, 89, 7065. RECEIVED January 6, 1993

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

Chapter 9

UV Photodissociation and Energy-Selective Ionization of Organometallic Compounds in a Molecular Beam

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch009

Jeffrey A. Bartz, Terence M. Barnhart, Douglas B. Galloway, L. Gregory Huey, Thomas Glenewinkel-Meyer, Robert J. McMahon, and F. Fleming Crim Department of Chemistry, University of Wisconsin—Madison, Madison, WI 53706

This chapter introduces the application of molecular beams, ultraviolet lasers, energy-selective ionization, and time-of-flight mass spectrometry to the general study of the Laser Chemistry of Organometallics. It briefly contrasts energy-selective vacuum ultraviolet ionization and time-of-flight mass spectrometry with other laser methods used in investigating the photodecomposition of organometallic molecules. After a description of the experimental apparatus, a summary of results from photodissociation studies of 5

dicarbonyl(η-cyclopentadienyl)(1-propyl)iron appears. LASERS AND ORGANOMETALLIC MOLECULES Lasers can provide useful information about the primary photodissociation pathways of organometallic compounds. Techniques such as chemical trapping, transient infrared spectroscopy, and mass spectrometry explore the photolyses of organometallic molecules in a "solvent-free" environment. These gas phase investigations generate information on energy disposal andfragmentationpathways for volatile organometallic systems. A notable example of early work on the gas phase photochemistry of organometallic molecules is that of Yardley and coworkers (1,2). They used PF to trap the fragmentsfromthe 248-nm photodissociation of Fe(CO). They found a distribution of coordinatively unsaturated iron carbonyl fragments, but, because collisions between different iron carbonylfragmentsrandomize the extent of unsaturation, they did not unambiguously determine the detailed distribution of the initial fragments. Direct spectroscopic methods can also detect the primary photodissociationfragmentsfromorganometallic compounds. Weitz and coworkers studied the photodissociation of Fe(CO) bytime-resolvedinfrared spectroscopy (TRIS) (3,4) and observed the CO stretchingfrequenciesoffragmentsformed by ultraviolet (UV) laser photolysis. They detected the same fragments trapped by Yardley and coworkers (1,2). Another method for detecting productsfromthe laser photodissociation of organometallic molecules is mass spectrometry using electron impact or multiphoton 3

5

5

0097-6156/93/0530-0116$06.00/0 © 1993 American Chemical Society In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

9.

BARTZ E T AL.

UV Photodissociation and Ionization

ionization. The ions observed in a mass spectrum help identify the primary photodissociation pathways of the molecules in a collision-free environment. As one example, Mikami et al. (5) used a molecular beam, multiphoton ionization (MPI), and time-of-flight mass spectrometry (TOFMS) to detect metal atoms and other fragments from the photodissociation of a series of metal acetylacetonates. Multiphoton ionization is a useful technique for identifying products, although many of the species come from secondary photolyses. Thus, the observed masses reflect more than just the primary photodissociation event. A more conventional ionization method, electron impact, can detect all products from a photodissociation of an organometallic compound. For example, Vernon and coworkers studied the 248-nm photodissociation of Fe(CO) (6), Z n ( C H ) (7) and M ( C O ) (8) (M = Cr, Mo, W) by electron impact mass spectrometry. The typical electron energies used for mass spectrometry both ionize the primary photodissociation products and extensively fragment those same ions. We have implemented an energy-selective method that avoids many of the complications of secondary dissociations in the ionization step of mass spectrometry detection. We use vacuum ultraviolet (VUV) ionization and TOFMS to identify the primary fragments from the photodissociation of a variety of molecules, including organometallic compounds. This method is energy-selective as it ionizes only the chemical species with ionization potentials of 9.9 eV or lower, which for most of them is near their ionization threshold. Because these ions have little excess energy from the V U V photon, they undergo secondary dissociations less extensively compared to those formed by MPI or electron impact ionization. Our approach is similar to the single photon ionization work of Van Bramer and Johnston (9), although we generate our V U V light differently. V U V ionization coupled with U V laser photodissociation and TOFMS is a valuable tool for determining decomposition pathways in a collision-free environment and is a means of identifying the important intermediates in chemical vapor deposition or in condensed phase photolysis. 5

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch009

117

2

5

2

6

METHOD Figure 1 shows a portion of our experimental molecular beam apparatus, which has two differentially pumped chambers separated by a skimmer. The organometallic compound resides in a sample holder, inside the molecular beam source chamber, approximately 5 cm behind the nozzle of the pulsed valve. Thermal coaxial cable heats this holder, which has an inlet for the external carrier gas, He, and an outlet that connects to the nozzle. Typically, the source chamber operates at a temperature of 80 °C and the carrier gas pressure is about 300 torr. The nozzle pulses at 20 Hz and produces 1 us-long bursts of gas. The valve sits in an aluminum holder, wrapped in thermal coaxial cable. Typically, the pulsed valve operates at 90 °C, a higher temperature than the sample holder, to reduce condensation or sublimation of the organometallic compound in the nozzle assembly. The seeded carrier gas travels out of the nozzle and expands, producing a molecular beam that passes through a skimmer and travels into the interaction region, about 3 cm from the nozzle. In the interaction region, 280-nm U V light crosses the beam of organometallic molecules and photolyzes it. After a short time delay (0-1000 ns), 125-nm V U V light ionize the resulting photofragments. An extraction field accelerates the ions into the field-free region of a TOFMS where they separate by mass. To remove any contributions from either the U V or V U V photons alone, the mass spectrum presented later is the result of subtracting the UV-only and VUV-only signal contributions from the signal acquired by the method described above. In general, these background contributions are negligible in the mass region presented.

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

O

M

o

o

o

O

H

w Pd o K W g

00

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch009

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

9.

BARTZETAL.

119

UV Photodissociation and Ionization

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch009

LASER W A V E L E N G T H S AND ENERGETIC CONSIDERATIONS The apparatus uses two separate Nd:YAG-pumped dye laser systems. The first produces U V light for photolysis (280 nm) by frequency-doubling the dye laser light. The second generates V U V light by four-wave mixing in mercury. Light from the second dye laser (625 nm) passes through a K D P crystal where a portion of the beam doubles in frequency. The two wavelengths (625 nm + 312.5 nm) travel co-linearly through a lens and into a Hg heat pipe where the combination of two 312.5-nm U V photons and one 625-nm visible photon produces a single 125-nm (9.9 eV) V U V photon through resonant four-wave mixing. The details of the heat pipe design appear elsewhere (10). As with electron impact and multiphoton ionization, small ionization cross-sections and secondary dissociations can obscure the pathways we hope to observe. Furthermore, the energy of the V U V photons limits the fragments detected to those with ionization potentials less than 9.9 eV. Published ionization potentials (11,12) and previous results from the photodissociation of 1-nitropropane (10) in this laboratory guide our assignment of primary photodissociation fragments. In the photolysis of ( n ^ 5 H ) F e ( C O ) ( C H C H O H ) described below, 9.9-eV photons ionize all the products, excluding CO, which has an ionization potential (14.0 eV) (11) greater than the energy of the V U V ionization photons (9.9 eV). 5

5

2

2

2

3

5

(ri -C5H5)Fe(CO)2(CH CH CH3) R E S U L T S 2

2

5

Solution phase photodissociation of (r| -C5H5)Fe(CO) (alkyl) 1 produces alkanes (13) and alkenes (14) (Scheme 1). The production of alkanes apparently 2

Scheme 1

Fe o c " I H

> -

R

begins with Fe-alkyl bond homolysis, and the resulting alkyl radical 3 abstracts a hydrogen atom from solvent to yield an alkane 4. The proposed primary step in the production of alkenes involves the photochemical generation of a vacant coordination site, either through C O loss to give 5 or through ring slippage. Subsequent insertion of the metal into a P-C-H bond (P-hydride insertion) produces a metal-alkene hydride complex 6. Alkane production occurs most readily in fluid solution, where geminate

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

120

LASER CHEMISTRY O F ORGANOMETALLICS

propyl C3H7

FeC H 5

FeC H

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch009

5

40

60

120

100

80

6

5

Mass 5

Figure 2. TOF mass spectrum of (7? -C H )Fe(CO) (CH CH CH ) following 280-nm photolysis. 5

5

2

2

2

3

radical pairs diffuse apart rapidly (13,15,16), while alkene production dominates in matrices at low temperature (14,17). The partitioning between alkane and alkene pathways is unknown eitKer in solution or in the gas phase. The measurements described here provide direct evidence for both pathways occurring in the gas phase. Figure 2 shows the mass spectrum obtained upon photodissociation of (rj C5H5)Fe(CO)2(CH CH2CH ) (1, R=CH ) at 280 nm. The presence of propyl and propene as primary photoproducts, following the absorption of one U V photon, provides strong evidence for the two separate photodissociation pathways. The products containing iron, FeCsHs from Fe-R bond homolysis and FeCsH6 from (3hydride elimination, also one-photon products, are an additional indication of two decomposition channels. Moreover, the products C5H5 and C5H6, which come from multiphoton dissociation, are also consistent with those photopathways. (To determine the number of photons required to produce any feature in the mass spetrum we measure the variation in product yield as a function of laser power.) The number of carbonyls attached to any Fe-containing intermediates following 280-nm photolysis remains unclear from our data. The detected FeCsHs and FeCsH6 ions may not correspond to the residual Fe-containing species following alkyl loss or alkene production because of excess energy from the ionizing V U V photon. With the absorption of a 9.9-eV photon, secondary loss of C O may occur for any product containing Fe-CO bonds. For example, the parent molecule 1, with an ionization potential between 7 and 8 eV (72), dissociates both carbonyls following V U V ionization. Despite these secondary dissociations, two distinct primary photodissociation pathways appear from the absorption of a 280-nm photon: Fe-alkyl bond homolysis and direct production of alkenes. These results agree qualitatively with those of condensed phase experiments (18) which produce alkenes (14,17) and, to a lesser extent, alkyl radicals (14-16). We are preparing a more detailed analysis of these data. 5

2

3

3

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

9.

BARTZ ET AL.

UV Plwtodissociation and Ionization

121

CONCLUSIONS The laser chemistry of organometallic compounds in a molecular beam provides clues to the mechanisms for photodecomposition of metal-containing species. The combination of U V photolysis, single-photon V U V ionization, and TOFMS, can observe many species that may also appear in solution or in laser-assisted chemical vapor deposition. In the case of dicarbonyl(r| -cyclopentadienyl)(l-propyl)iron, the measurements clearly show that both alkyl radical and alkene production are primary processes. 5

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch009

ACKNOWLEDGMENTS J.A.B. wishes to thank the U.S. Department of Education for fellowship support, and T.G.M. gratefully acknowledges the support of the Alexander von Humboldt Foundation by a Feodor-Lynen Fellowship. The Army Research Office and the National Science Foundation support this work.

Literature Cited (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)

Nathanson, G.; Gitlin, B.; Rosan, A. M.; Yardley, J. T. J. Chem. Phys. 1981, 74, 361-369. Yardley, J. T.; Gitlin, B.; Nathanson, G.; Rosan, A. M. J. Chem. Phys. 1981, 74, 370-378. Ouderkirk, A. J.; Weitz, E. J. Chem. Phys. 1983, 79, 1089-1091. Ryther, R. J.; Weitz, E. J. Phys. Chem. 1992, 96, 2561-2567. Mikami, N.; Ohki, R.; Kido, H. Chem. Phys. 1990, 141, 431-440. Ray, U.; Brandow, S. L.; Bandukwalla, G.; Venkataraman, B. K.; Zhang, Z.; Vernon, M. J. Chem. Phys. 1988, 89, 4092-4101. Hou, H.; Zhang, Z.; Ray, U.; Vernon, M. J. Chem. Phys. 1990, 92, 1728-1746. Venkataraman, B.; Hou, H.; Zhang, Z.; Chen, S.; Bandukwalla, G.; Vernon, M. J. Chem. Phys. 1990, 92, 5338-5362. Van Bramer, S. E.; Johnston, M. V. Anal. Chem. 1990, 62, 2639-2643. Huey, L. G. Ph. D. Thesis, University of Wisconsin - Madison, 1992. Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.; Mallard, W. G. Gas-Phase Ion and Neutral Thermochemistry; Supplement 1 ed.; American Chemical Society and American Institute of Physics: New

(12) (13) (14) (15) (16) (17) (18)

York, 1988; Vol. 17. Lichtenberger, D. L.; Fenske, R. F. J. Am. Chem. Soc. 1976, 98, 50-63. Alt, H. G.; Herberhold, M.; Rausch, M. D.; Edwards, B. H. Z. Naturforsch. 1979, 34b, 1070-1077. Kazlauskas, R. J.; Wrighton, M. S. Organometallics 1982, 1, 602-611. Blaha, J. P.; Wrighton, M. S.J.Am. Chem. Soc. 1985, 107, 2694-2702. Hudson, A.; Lappert, M. F.; Lednor, P. W.; MacQuitty, J. J.; Nicholson, B. D. J. Chem. Soc. Dalton Trans. 1981, 2159-2162. Mahmoud, K. A.; Rest, A. J.; Alt, H. G. J. Chem. Soc. Dalton. Trans. 198 1365. Pourreau, D. B.; Geoffroy, G. L. In Advances in Organometallic Chemistry; F. G. A. Stone and R. West, Ed.; Academic Press: New York, 1985; Vol. 24; pp

249-352.

RECEIVED

January 25, 1993

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

Chapter 10

Buffer-Gas Quenching as a Dynamical Probe of Organometallic Photodissociation Processes 1

Rong Zhang and Robert L . Jackson

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch010

I B M Research Division, Almaden Research Center, 650 Harry Road, San Jose, CA 95120

We have developed an indirect method for probing the photodissociation dynamics of organometallic molecules. This method relies on the sequential multiple-bond-scission processes that typically occur in the photodissociation of organometallics, even upon absorption of a single photon. Photodissociation is carried out in the presence of a buffer gas and the yields of both the primary and secondary products are determined as a function of buffer-gas pressure. The resulting "quenching curve" gives the pressuredependent probability that either the primary or secondary dissociation processes are quenched via collisional stabilization of the reactant. The quenching data are then fit using a time-dependent master equation to describe the competition between reaction and stabilization. From the fitted data, we can readily determine if the primary photodissociation process occurs in a statistical manner, and if vibrational energy is partitioned statistically to the primary photoproducts. The quenching technique is illustrated for the photodissociation of two distinctly different classes of organometallic molecules: the dialkyl zincs and ferrocene.

Experimental studies of photodissociation dynamics have contributed extensively to our understanding of chemical reactions at the molecular level (1). The development of key experimental techniques over the years, including molecular beams, laserinduced fluorescence (LIF), and resonance-enhanced multiphoton ionization (REMPI), has enabled researchers to determine the translational, rotational, and vibrational energy distributions of products formed upon photodissociation of small and intermediate-sized molecules (1). By comparing such distributions with theoretical predictions, it is possible to learn a great deal about the dynamics of many different classes of photoinitiated reaction processes. Studies of the photodissociation dynamics of large molecules, including organic and organometallic species, have not kept pace with studies of small molecules, primarily because of experimental difficulties. For example, it is difficult to resolve Current address: SRI International, Menlo Park, CA 94025 0097-6156/93/0530-0122$06.00/0 © 1993 American Chemical Society In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch010

10.

ZHANG AND JACKSON

Buffer-Gas Quenching as a Dynamical Probe

and interpret spectroscopic detail at the rotational level, and sometimes even the vibrational level, for molecules composed of more than a few atoms, so experimental techniques such as LIF and REMPI that seek to determine product state distributions spectroscopically are rarely applicable to large product molecules. Also, organometallics present a unique problem in molecular-beam experiments because of their tendency to undergo multiple bond-cleavages upon absorption of only a single photon (2). Using a crossed laser-molecular beam apparatus, one can determine photoproduct translational energy distributions by monitoring the time required for products to travel from the point of creation to the detector, which is usually a mass spectrometer. If the products spontaneously dissociate on the way to the detector, however, there is no unique point of product origin, making it difficult to determine product translational energy distributions uniquely (3). Also, because organometallics tend to fragment extensively upon electron impact, it can be difficult to tell whether a given product ion results from photodissociation, electron-impactinduced fragmentation in the mass spectrometer, or both (3,4). Because of these experimental limitations, we have begun to apply an indirect method to examine the photoreaction dynamics of organometallics in the gas phase (5-7). We photodissociate an organometallic molecule with a pulsed laser and determine the yields of the various products, including those formed in spontaneous secondary fragmentation reactions. Photodissociation is performed in the presence of a buffer gas in order to stabilize the parent molecule or the primary photoproducts before they can undergo secondary dissociation. The probability of stabilization vs. dissociation is determined at several pressures. The resulting quenching data are then fit using a time-dependent master equation (8), treating the spontaneous secondary dissociation processes, or the primary photodissociation process itself, as statisticallydriven reactions. By comparing the calculated and experimental quenching curves, we can learn about the dynamics of the primary photodissociation process without direcdy measuring product state distributions. While our method is suitable only for photoprocesses where the primary photoproducts undergo spontaneous secondary dissociation, this is rarely a limitation for organometallics, since such secondary dissociation processes are the rule, rather than the exception. We demonstrate how the quenching technique can be used to probe the photodissociaton dynamics of organometallic molecules, presenting results obtained upon photodissociation of two distinctly different types of organometallic compounds: the dialkyl zincs (5,6) and ferrocene (7). Experimental Details of the experimental apparatus and product detection methods have been described elsewhere (6), so we will only provide a brief outline here. A schematic of the experimental appratus is shown in Fig. 1. Photodissociation was conducted in a stainless steel flow cell using a pulsed excimer laser as a light source. Product detection was accomplished via LIF driven by a counter-propagating N d : Y A G pumped dye laser tuned to specific product bands. Fluorescence was collected at normal incidence, dispersed with a monochromator, and detected using a photomultplier tube or diode array. The fluorescence intensities and relative yields are corrected for variations in the intensity of the excimer and dye laser beams during a photolysis run, as well as for pressure effects on spectroscopic linewidths and fluorescence lifetimes. The organometallic compound was contained within a stainless steel vessel and was delivered into the cell using He as a carrier gas. A separate carrier gas line was also attached to the cell, allowing the organometallic reactant and He partial pressures to be controlled separately. Flow rates were controlled accurately using mass flow controllers. Caution: the dialkyl zinc compounds flame spontaneously upon exposure to the atmosphere.

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

123

124

LASER CHEMISTRY OF ORGANOMETALLICS

Photodissociation of the Dialkyl Zincs We begin by briefly reviewing prior photochemical and spectroscopic work on the dialkyl zincs as well as the dialkyls of the other Group 12 metals, Cd and Hg. The U V spectra of these metal alkyls show two broad, overlapping bands in the deep U V (9,10). The lower energy band has little or no discernible structure, while the higher energy band shows diffuse vibrational structure. The linewidths are consistent with a photodissociation lifetime on the order of 50 fs (10). The photodissociation process yields metal atoms via a stepwise process (77): R M RM "

—

2

• •

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch010

1

RM+ + R R + M

(1) (2)

where R denotes an alkyl radical and * denotes vibrational excitation. For M=Zn, the vibrationally hot monoalkyl intermediate, RZn*, must retain about half of the energy partitioned to products in the 248-nm photodissociation process to undergo spontaneous secondary dissociation to Zn atom (see Table I). Table I: Bond dissociation energies

a

Molecule

DH°i (kcal/mole)

D H ° (kcal/mole)

(CH ) Znb

63.7 ± 1.5

24.5 ± 4.0

(C H ) Znb

52.4 ± 2.0

22.0 ±4.2

(n-C H ) ZnC

52.4 ± 2.0

22.0 ± 4.2

3

2

2

5

3

2

7

2

FeCp

91 ± 3

2

2

67 ± 5 ^

d

SOURCE: Adapted from Ref. 6. a

D H ^ i represents the dissociation energy of the first metal-ligand bond, while represents the dissociation energy of the second metal-ligand bond. Jackson, R.L. Chem. Phys. Lett. 1989,163, 315. Bond dissociation energies in ( n - C H / ) Z n are assumed to be equal to those in (C H5) Zn. See Benson, S.W.; Francis, J.T.; Tsokis T.T. J. Phys. Chem. 1988, 92, 4515. Ref. 30. Puttemans, J.P.; Smith, G.P.; Golden, D . M . / . Phys. Chem. 1990, 94, 3226. D

c

3

2

2

2

d

e

In our work (5,6), we detect the yield of Zn atom and of the monoalkyl zinc intermediate via LIF upon photodissociation of the dialkyl zinc in the presence of He buffer gas. The yields of Zn and C H Z n vs. He pressure are shown in Fig. 2 for photodissociation of ( C H ) Z n . The yield of Zn falls, while the yield of C H Z n rises with increasing buffer-gas pressure because the secondary spontaneous dissociation step [Eqn (2)] is quenched with increasing efficiency as the buffer gas pressure rises. The yield of Zn falls from -95% at very low pressure to -50% at 400 torr. The yields shown in Fig. 2 are determined by normalizing the LIF data to a total yield (Zn + C H Z n ) of unity at all pressures. Similar results are obtained upon photodissociation of ( C H 5 ) Z n and (nC H;) Zn, shown in Fig. 3. Again, the yield of Zn falls with increasing He pressure. Unfortunately, we could not obtain complementary data on the yield of the 3

3

2

3

3

2

2

a s

3

2

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

10.

Buffer-Gas Quenching as a Dynamical Probe

ZHANG AND JACKSON

Pulse Generator

Delay Generator Diffusion Pump

Fluorescence Viewport

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch010

Dye Laser

M a s s Flow Controller

Mass Flow Controller

Buffer Gas

R Zn/He Mixture 2

Figure 1: Experimental apparatus (Reproduced with permission from Ref. 6. Copyright 1992 American Institute of Physics).

2 5 CD

0

100

200

300

400

Pressure (Torr)

Figure 2: Yield of Zn (A) and C H Z n (•) obtained upon photodissociation of (CH3)2Zn vs. He buffer gas pressure. The data are normalized to an average total yield of 1.0 (O), represented by the dotted line (Reproduced with permission from Ref. 6. Copyright 1992 American Institute of Physics). 3

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

125

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch010

126

LASER CHEMISTRY OF ORGANOMETALLICS

o I 0

1

I

100

I

I

I

200

Pressure

I

300

1

1

1

400

(torr)

Figure 3: Yield (•) of Zn obtained upon photodissociation of (C2H5)2Zn (top) and (n-C3H7)2Zn (bottom) vs. He pressure. The data are normalized to an extrapolated (Stern-Volmer) yield of unity at zero pressure. The solid line represents the best fit obtained from master equation calculations (see text). The dotted line represents the fit obtained assuming that the vibrational energy distribution is a Gaussian function with a width (FWHM) of 3000 cm"*. The dashed line represents the fit obtained assuming that the vibrational energy distribution is given by phase-space theory (Reproduced with permission from Ref. 6. Copyright 1992 American Institute of Physics).

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

10. Z H A N G AND JACKSON

Buffer-Gas Quenching as a Dynamical Probe n

a n

monoalkyl zinc for (C2^s)2^ ^ (n-C3H7)2Zn, since no fluorescence signals attributable to C2H5Zn and n-C3H7Zn were observed. As discussed in Ref. 6, the lack of fluorescence from these species can be understood from symmetry arguments. Still, the declining yield of Zn with increasing He pressure confirms the two-step photodissociation mechanism for (C2H5)2Zn and (n-C3H7)2Zn, and confirms that the C2H5Zn and n-C3H7Zn intermediates are stabilized with increasing probability as the He pressure rises.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch010

Time-Dependent Master Equation Calculations. To determine the nascent vibrational energy distribution of the monoalkyl zinc intermediate from the quenching data shown in Figs. 2 and 3, we use a fitting procedure based on a time-dependent master equation formalism. The following is a brief outline of this procedure, since it has been described in detail elsewhere (8). Denoting the monoalkyl zinc species as A , the appropriate form of the time-dependent master equation is:

s

"

di

= A(0)(|)(E)dE +coJP(E,E')A(E ,t)dE' - coA(E,t) - k(E)A(E,t). ,

(3)

0 The first term on the right-hand side of Eqn (3) defines the concentration and vibrational energy distribution of A created by photodissociation at t=0, the second and third terms define the rate of collisional energy transfer between the monoalkyl zinc and the buffer gas, while the fourth term defines the rate of spontaneous dissociation of the monoalkyl zinc. The meaning of the variables and their values are given below: (|)(E)dE is the probability that photodissociation yields a monoalkyl zinc molecule with internal energy between E and E + dE. This function is the differential form of the nascent vibrational energy distribution for the monoalkyl zinc photoproduct, and is treated as an adjustable function in order to fit the quenching data. co is the rate of collisions between A and the buffer gas. We employed LennardJones collision rates, which were determined as described in Ref. 6. k(E) is the J-averaged (72) R R K M rate constant for dissociation of the monoalkyl zinc intermediate. The R R K M calculations are described in detail in Ref. 6. P(E',E) is the normalized probability that a single collision with He transforms A(E) into A(E'). Theories of collisional vibrational energy transfer are not yet sufficiently well developed for polyatomic molecules to yield a general functional form for P(E',E), so a suitable mathematical model must be specified. A common model, and the one we used in this work, is the exponential-down model (13): P(E',E)=^exp[-^|)

E > E',

(4)

where N(E) is a normalization factor and a(E) is a parameter with units of energy that denotes the average quantity of vibrational energy transferred per collision. The value of a(E) can be determined from direct measurements, as have been performed on several polyatomic molecules (74), but such data are not available for the monoalkyl zincs. Semi-quantitative data on the rate of thermalization of the hot CH3Zn photoproduct are available from our LIF experiments, however, as described in Ref. 6. The value of oc(E) for each calculation was chosen to be consistent with the measured rate of thermalization. Analogous data are not available for C2H5Zn and nC3H7Zn, but recent studies indicate that the average quantity of vibrational energy removed per collision by a given buffer gas does not vary greatly from one molecule to the next (75). Hence, we used the thermalization rate data for CH3Zn to determine a(E) for all the monoalkyl zincs. In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

127

128

LASER CHEMISTRY O F ORGANOMETALLICS

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch010

The goal of our fitting procedure is to determine the function φ(Ε) that provides the best fit to the quenching data for all three dialkyl zinc compounds. It is not possible to determine φ(Ε) uniquely, since many functions could conceivably fit the data. We restricted the set of trial φ(Ε) functions to continuous functions, such as a Gaussian, Boltzmann, or Poisson distribution. Use of these model functions will give a reasonably accurate picture of the peak energy and the width of the nascent internal energy distribution for the monoalkyl zinc photoproducts. Fitting Results for (C2H5)2Zn and ( η ^ Η γ ^ Ζ η . We found that good fits to the quenching data for (C2H5)2Zn and ( η ^ Η γ ^ Ζ η could be obtained only by assuming that the nascent vibrational energy distribution of the monoalkyl zinc photoproduct was narrow and quite hot, i.e. highly non-statistical. For (C2H5)2Zn, a good fit to the quenching data is obtained by representing φ(Ε) as a Gaussian function with a fullwidth at half-maximum of 500 cm"! centered at 7600 cm"*. For (η-ϋβΗγ^Ζη, a good fit to the quenching data is obtained by representing φ(Ε) as a Gaussian function with a full-width at half-maximum of 1250 cm"* centered at 9200 c m . Figure 3 shows the fits to the quenching data obtained with these choices of φ(Ε). The hot, narrow vibrational distribution functions that provide the best fit to the quenching data for (C2H5)2Zn and ( η ^ Η γ ^ Ζ η may be compared with the distributions derived from phase-space theory (PST) (76) or separate statistical ensemble (SSE) theory (77). Each theory has been used successfully to compute photoproduct rotational and vibrational distributions for photodissociation processes that follow a statistical dissociation pathway. The vibrational energy distribution functions predicted by each theory for the monoalkyl zinc photoproducts are approximately Gaussian in shape, but are quite broad, as shown in Table II. Fig. 3 shows that these broad distribution functions provide a very poor fit to the quenching data for both (C2H5)2Zn and ( η ^ Η γ ^ Ζ η . - 1

Table II. Energy maximum and width (FWHM) of the vibrational energy distributions computed via PST and SSE theory for the monoalkyl zincs PST Molecule

SSE 1

Maximum (cm ) Width (cm" ) Maximum (cm ) Width (cm" ) -1

-1

1

CH Zn

7800

8200

9200

7500

C H Zn

10800

7600

11200

7800

n-C3H Zn

11100

7000

9000

7600

3

2

5

7

SOURCE: Reprinted with permission from Ref. 6. To ensure that our fitting results are accurate, we adjusted the values of k(E) and ωΡ(Ε',Ε) in order to force a fit using the monoalkyl zinc vibrational energy distributions predicted by phase-space theory. A reasonable fit can be obtained, but only by making huge changes to one or both rate constants. For (C2H5)2Zn, the R R K M rate constants would have to be reduced by a factor of -100 in order to force a reasonable fit. Alternatively, the calculated collisional energy transfer rates would have to be increased by a factor of -100. For ( η ^ Η γ ^ Ζ η , the errors would also have to be quite large (-35 x). There is no reason to believe that our calculations have such large errors.

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

10.

Z H A N G AND JACKSON

Buffer-Gas Quenching as a Dynamical Probe

To rationalize the hot, narrow vibrational energy distribution found for the C2H5Zn and η-ΟβΗγΖη photoproducts, we examine the photodissociation pathways that are available in the dialkyl zincs. Fig. 4 shows a one-dimensional cross-section, taken along the breaking C-Zn bond, of the potential energy surfaces involved. The symmetry species assigned to each state are based on the highest molecular symmetry (C ) preserved throughout the course of the bond-breaking process. Excitation of the dialkyl zinc at 248 nm populates the lower-energy ^A' excited state, but this state cannot dissociate directly to ground-state products without crossing onto the repulsive 3A' surface or internally converting onto the ground-state potential surface. We can readily eliminate the internal conversion mechanism, since it is well-known that dissociation processes of this kind follow a statistical pathway (17-20). Hence, we conclude that photodissociation proceeds via crossover from the optically prepared ^A' exicted state to the repulsive 3A' state. Because bond-breaking occurs on the time scale of a single vibration, however, it is probably more precise to consider the optically prepared state to be an admixture of the ^A' excited state and the repulsive A ' state, with the 1 A state carrying most of the oscillator strength. Based on this picture of the potential surfaces involved in the photodissociation process, we propose the following explanation for the hot, narrow vibrational energy distribution of the monoalkyl zinc photoproduct. The dialkyl zinc molecule is placed on the repulsive 3 A' potential surface within a very narrow region of phase space defined by several factors, including the wavelength of the photon source, the FranckCondon matrix elements connecting the ground state and the ^A' excited state, and the spin-orbit matrix elements connecting the ^A' excited state with the repulsive A ' state. Because the accessible phase space is limited, the product vibrational energy distributions are much narrower than those expected from statistical theories. Also, both the alkyl radical and monoalkyl zinc fragments are born with a geometry nearly identical to their geometries within the ground-state dialkyl zinc. As discussed in Ref. 6, these geometries differ significantly from the equilibrium geometries for the separated fragment, so both fragments are formed with a relatively hot vibrational energy distribution.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch010

s

3

1

3

Fitting Results for (CH3)2Zn. In contrast to the good fits obtained for (C2H5>2Zn and (n-C3H7)2, a good fit could not be obtained for (CH3)2Zn. Neither a narrow Gaussian vibrational distribution function, such as those used successfully for (C2H5)2Zn and ( η ^ Η γ ^ Ζ η , nor the broader vibrational distribution functions predicted by PST or SSE could fit the quenching data. For all reasonable distribution functions, the effect of He on the spontaneous secondary dissociation probability for the CH3Zn photoproduct was much stronger than predicted by our calculations. To understand why the master equation procedure fails for (CH3)2Zn, we need only look at the R R K M calculations on the dissociation rate of the CH3Zn photoproduct. CH3Zn is a small molecule with a low barrier to dissociation. As a result, the vibrational state density of CH3Zn in the vicinity of the dissociation barrier is extremely low, ~10/cm~l. Experiments by Bauer and coworkers (27) indicate that R R K M theory fails to predict reaction rates in cases where the state density of the reactant is so low. The reason for this failure is that the rate of internal vibrational energy redistribution (IVR) is not fast enough at very low state densities to ensure that the reaction follows a purely statistical pathway. R R K M theory will thus predict a reaction rate that is much faster than the actual reaction rate, translating into calculated quenching probabilities that are much smaller than observed. Photodissociation of Ferrocene We now turn to a discussion of our work on the photodissociation of ferrocene (7) beginning with a brief review of prior work. Ferrocene has been irradiated numerous

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

129

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch010

130

LASER CHEMISTRY OF ORGANOMETALLICS

>

r

C-Zn

Figure 4: Cross section taken along the breaking C-Zn bond of the potential energy surfaces involved in photodissociation of the dialkyl zincs (Reproduced with permission from Ref. 6. Copyright 1992 American Institute of Physics).

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch010

10.

Z H A N G AND JACKSON

Buffer-Gas Quenching as a Dynamical Probe

times in condensed phases, and it is clear from published results that the quantum yield for photoreaction is immeasurably low (22). Excitation of ferrocene in the U V is apparently followed immediately by internal conversion to the ground state, since no fluorescence is observed and the phosphorescence quantum yield is very low (23). Internal conversion in condensed phases is of course followed very rapidly by energy transfer to the surrounding medium. Since the first Fe-Cp bond dissociation energy is quite large (see Table I), dissociation cannot compete with relaxation following absorption of a single photon. While the photostability of ferrocene in condensed phases is well-documented, it has been known since the 1956 flash photolysis work of Thrush (24) that photodissociation of ferrocene in the gas phase yields measurable quantities of cyclopentadienyl radical and iron atoms. Recent laser studies by several groups show that dissociation is a multiphoton process (25-28). Gas-phase multiphoton dissociation of ferrocene proceeds stepwise (25-28), with the monocyclopentadienyl iron species formed as an intermediate: Cp2Fe 1

CpFe "

nhv

*

CpFe * + Cp

(5)

>

Cp + Fe

(6)

1

There is disagreement over the dynamics of the primary dissociation process [Eqn (5)], however. Engelking and coworkers concluded from multiphoton ionization studies of the iron atom photoproduct that multiphoton dissociation of ferrocene at wavelengths

"ω

ο ι 0

I

I

I

40

I

I

80

I

I

120

I

1

160

Pressure (Torr)

Figure 5: Relative yields as a function of He pressure of cyclopentadienyl radical ( · ) and iron atom (O) produced in the two-photon dissociation of ferrocene at 248 nm.

ο I 0

1

I

40

I

I

I

80

I

120

I

1

1

160

Pressure (Torr)

Figure 6: Fit of the cyclopentadienyl yield as a function of He pressure ( · ) to the yield calculated (solid line) using Eqn (3), assuming that the primary photodissociation process is statistical.

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

133

134

LASER CHEMISTRY OF ORGANOMETALLICS

in k(E) is not possible. Our estimating procedure for P(E',E) is undoubtedly subject to some error, but since we believe that our estimated value lies below than the actual value, it is extremely unlikely that P(E',E) could actually be overestimated by a factor of seven. Our calculations do not reveal the actual mechanism for the primary photodissociation step, but they show conclusively that it occurs much faster than the statistically predicted rate. Such results would be expected for a photodissociation process that involves predissociation or direct excitation to a repulsive potential surface. Our results thus support the conclusions of Engelking and coworkers (27) that dissociation occurs via the direct process.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch010

Literature Cited

1. For a general discussion, see "Advances in Gas-Phase Photochemistry Kinetics, Molecular Photodissociation Dynamics", Ashfold, M.N.R.; J.E., Eds.; Royal Society: London, 1987. 2. For example, see Weitz, E. J. Phys. Chem. 1987, 91, 3945. 3. Venkataraman, B.; Hou, H.; Zhang, Z.; Chen, S.; Bandukwalla, G.; Vernon, M. J. Chem. Phys. 1990, 92, 5338. 4. Ray, U.; Brandow, S.L.; Bandukwalla, G.; Venkataraman, B.K.; Zhang, Z.; Vernon, M. J. Chem. Phys. 1988, 89, 4092. 5. Jackson, R.L. J. Chem. Phys. 1990, 92, 807. 6. Jackson, R.L. J. Chem. Phys. 1992, 95, 5938. 7. Zhang, R.; Jackson, R. L., manuscript in preparation. 8. Jackson, R.L. Chem. Phys. 1991,157,315. 9. Chen, C.J.; Osgood, R.M. J. Chem. Phys. 1984, 81, 327. 10. Amirav, Α.; Penner, Α.; Bersohn, R.J.Chem. Phys. 1989, 90, 5232. 11. Jonah, C.; chandra, P.; Bersohn, R. J. Chem. Phys. 1971, 55, 1903. 12. Smith, S.C.; Gilbert, R.G. Int. J. Chem. Kinet. 1988, 20, 307. 13. For example, see Gilbert, R.G.; Luther, K.; Troe, J. Ber. Bunsenges. Phys. Ch 1983, 87, 169. 14. Oref, I.; Tardy, D.C. Chem. Rev. 1990, 90, 1407. 15. Damm, M.; Deckert, F.; Hippler, H.; Troe, J. J. Phys. Chem. 1991, 95, 2005. 16. Pechukas, P.; Light, J.C. J. Chem. Phys. 1965, 42, 3281. 17. Wittig, C.; Nadler, I.; Reisler, H.; Noble, M.; Catanzarite, J.; Radhakrishnan, G. J. Chem. Phys. 1985, 83, 5581. 18. Hippler, H.; Luther, K.; Troe, J.; Wendelken, H.J. J. Chem. Phys. 1983, 79, 239. 19. Klippenstein, S.J. Chem. Phys. Lett. 1990, 170, 71. 20. Chen, I.; Green, Jr., W.H.; Moore, C.B. J. Chem. Phys. 1988, 89, 314. 21. Borchardt, D.B.; Bauer, S.H. J. Chem. Phys. 1986, 85, 4980 and references cited therein. 22. For example, see Borrell, P.; Henderson, E. Inorg. Chim. Acta 1975,12,215. 23. Smith, J.J.; Meyer, B. J. Chem. Phys. 1968, 48, 5436. 24. Thrush, B.A. Nature 1956, 178, 155. 25. Leutwyler, S.; Even, U.; Jortner, J.J.Phys. Chem. 1981, 85, 3026. 26. Liou, H.T.; Ono, Y.; Engelking, P.C.; Moseley, J.T. J. Phys. Chem. 1986, 90, 2888. 27. Liou, H.T.; Engelking, P.C.; Ono, Y.; Moseley, J.T. J. Phys. Chem. 1986, 90, 2892. 28. Ray, U.; Hou, H.Q.; Zhang, Z.; Schwarz, W.; Vernon, M. J. Chem. Phys. 1989, 90, 4248. 29. Nelson, H.H.; Pasternack, L.; McDonald, J.R. Chem. Phys. 1983, 74, 227. 30. Lewis, K.E.; Smith, G.P. J. Am. Chem. Soc. 1984, 106, 4650. RECEIVED

February 10, 1993

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

Chapter 11

Photoionization and Photodissociation Dynamics for Small Clusters of Refractory Metals Time-Resolved Thermionic Emission 1

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch011

Andreas Amrein , Bruce A. Collings, David M. Rayner, and Peter A. Hackett Steacie Institute for Molecular Sciences, National Research Council of Canada, 100 Sussex Drive, Ottawa, Ontario K1A 0R6, Canada

The chemistry of "naked" metal clusters, as revealed by molecular beam studies, is reviewed in the general context of chemistry at metal atom sites. Special emphasis is given to the chemistry of niobium clusters and their derivatives. The recent observation of delayed ionization from multiphoton excited, strongly bound clusters is highlighted. Recently, we have used this novel form of molecular ionization process to develop a reliable and calibrated thermometer for the vibrational temperature of metal cluster beams. This advance has meant that temperature may be used as a well defined variable in metal cluster beam experiments. This will be helpful for the development of "molecular" surface science using derivitized metal cluster beams. The agglomeration of metal particles is a common feature of the laser chemistry of organometallics. Such particles may be desirable intermediates in metal deposition or laser writing processes. They may also provide particularly effective supports or active media for heterogeneous catalysis. The physical or chemical properties, for instance the porosity, of these catalysts may be varied over wide ranges in materials which have well defined particle size distributions on the length scale of angstroms i.e. "cluster assembled" materials (1). Finally, small metal particles may be the undesired end products of too vigorous a laser photochemical reaction. The properties of the smallest of metal particles, the metal clusters, are as interesting as their occurrence is ubiquitous. A small fraction of these properties is reviewed in this contribution. Current address: Technikum Winterhur, Postfach 805, CH-8401, Switzerland

0097-6156/93/0530-0136$06.00/0 Published 1993 American Chemical Society In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

11.

AMREIN E T AL.

137

Small Clusters of Refractory Metals

The Role of "Naked" Metal Clusters in Metal-Ligand Chemistry The chemistry and photochemistry of "naked" metal clusters is an important area of the chemistry occurring at metal centres. General features of this chemistry are outlined in the matrix represented in Figure 1 in which the number of metal atoms and the number of ligands present in a complex M L provide a logical basis with which to categorize areas of chemistry ranging from surface science and organometallic chemistry through to metal atom chemistry. Within this context, the roles (i) of complexes, M L , between a single ligand and a metal atom, as prototypical organometallic systems, (ii) of coordinatively unsaturated organometallic species, M L , as reactive intermediates in organometallic catalytic cycles, and (iii) of single ligands on extended metal surfaces, M ^ L , as reactive intermediates in surface catalyzed reactions, are linked with those of "naked" metal clusters, M . Moscovits discussed these links in his prescient review of the field in 1979, emphasizing the essential differences in the ability of metal atoms in these various environments to accommodate ligands through electronic or geometrical reorganization (2). It is apparent that small metal clusters may act as of finite models for the extended phase of metallic materials since molecular processes observed in clusters may have analogues in bulk matter. By observing the dynamics of molecular processes in clusters one can gain a detailed understanding, using the language of molecular science, of processes occurring in the extended phase. x

y

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch011

X

x

ML •

M L 3

2

ML, +L ML

l

•

2

ih -11

Μ,

···

+M

·

M

n

!h BULK

—

Figure 1: Chemistry at metal centres.

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

138

LASER CHEMISTRY OF ORGANOMETALLICS

If we are resolved to use small derivitized clusters, M L , as simple models for examining chemistry occurring on extended metal surfaces, then we are required to define the properties, particularly the electronic and geometrical structural properties, of the "naked" metal cluster, M . A working definition of the term cluster is "a collection of atoms for which only the number of constituents is known". Once we know any molecular property of a cluster, the term cluster becomes inappropriate and the item becomes a molecule. A cluster of sixty carbon atoms was an interesting curiosity but aromatic C , with the soccer ball structure, is a far more compelling object. Our strategy is clear. We shall attempt to determine the molecular properties of small metal clusters and then we shall use the chemistry observed between these cluster molecules and ligands as simple molecular models for chemistry occurring in the extended state. In this presentation, we shall review progress to date, using molecular beam techniques, on the chemistry of refractory metal clusters particularly niobium. We shall highlight the recent observation of a new molecular ionization process observed in these clusters. This molecular process is equivalent to the property of thermionic emission in bulk metals. We shall show how we have been able to use this process to explore the vibrational energy content of metal clusters on metal cluster beams. This work has allowed us to define the vibrational temperature of metal cluster beams for the first time and opens the way to the use of vibrational temperature as a well defined variable for future studies of reactions of "naked" metal clusters. X

x

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch011

6 0

Chemistry of Niobium Clusters as Revealed by Molecular Beam Studies The chemistry of niobium clusters is interwoven with the general development of metal cluster beam science as niobium is an ideal element with which to demonstrate aspects of metal cluster beam technology. Laser vaporization of niobium easily leads to clusters of high nuclearity. Niobium has only one naturally occurring isotope and it is therefore an ideal subject for mass spectroscopy. Ionization Potentials for Niobium Clusters. Detailed threshold photoionization studies of niobium clusters have been made and ionization potentials for N b to N b have been reported (3,4)· The dependence of the ionization potential on cluster size is generally consistent with that expected from a consideration of image potential effects on the ionization energy of metal spheres (radius R) from which classical electrostatics would suggest equation 1 : 2

76

2

IP = W.F. + (l/2)e /R

(1).

A smooth decrease in ionization potential with increasing nuclearity is indeed observed. Thus, it would seem that the evolution of this particular property from molecule to bulk is well understood. However, as Moscovits has pointed out (5), this interpretation is too simplistic. Firstly, equation 1 is not followed exactly. In this size range a coefficient of 0.2 rather than 0.5 fits the data. Secondly, the work function of bulk niobium is highly dependent on the choice of crystal face (

M

(3)

r

hv, hv M

n

->

M

M

n

-»

M

M

L

n-i + n

(4)

n

->

M,

(5)

+

M

(6)

+

(7) #

n

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch011

#

Here, M * , represents an electronically excited cluster, M and M * represent clusters in which the excitation is distributed over all vibronic states acessible at the one and two photon energies respectively. Statistical calculations of the rates of reactions 6 and 7, based on assumed molecular properties of the relevant clusters, were in accord with the experimentally determined values (34). Calculations based upon size-corrected materials properties also gave reasonable agreement (34). An interesting result of the delayed ionization experiments was the behavior of derivitized clusters in which carbon or oxygen were substituted for niobium. Niobium oxide clusters, N b O , do not give delayed ionization and, because strongly bound NbO is an ideal leaving group, it was suggested that dissociation was favoured for these clusters. For niobium carbide clusters, N b . C , delayed ionization was observed. The ionization rate constant was independent of η but dependent on x. This intriguing result is explained by the fact that a statistical reaction rate constant is strongly dependent only upon the number of particles if the thermochemistry is similar, the total energy is equal and if all vibrational degrees of freedom are in their high temperature limit. These conditions are evidently satisfied by the niobium carbide system (34). These observations have recently been confirmed by ab initio calculations, using local density functional theory, of the bond dissociation and ionization energies of Nb . O and N b . C (Salahub, D. and Erickson, L., University of Montreal, Personal communication, 1992). The rate of reaction 3, k , has recently been investigated using autocorrelation measurements of the delayed ionization yield following excitation by 50 ps pulses of the third harmonic of a Nd-YAG laser. A lower limit of 2 χ 10 s was found for k (Collings, Β. Α., Rayner D. M . and Hackett P. Α., National Research Council of Canada, unpublished data). This observation supports the proposed mechanism in which it was assumed that k was fast. Detailed studies of the rates of delayed ionization have now been reported for tungsten (35), tantalum (36) and carbon clusters (37). n

x n

n

n

x

x

n

n

x

n

n

n

n

relaxa(jon

9

1

relaxation

relaxation

The Temperature Dependence of the Thermionic Emission Rate and The Vibrational Temperature of Metal Cluster Beams. Casual inspection of the mechanism outlined in equations 2-8 shows that the delayed ionization rate should depend upon the initial vibrational energy content of M . Recently experiments have been carried out which confirm this expectation n

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

11.

AMREIN E T AL.

Small Clusters of Refractory Metals

(Amrein, Α., Collings, Β. Α., Rayner D. M . and Hackett, P. Α., National Research Council of Canada, unpublished data). The experimental apparatus has been described previously (34). Briefly, metal clusters are formed in a laser vaporization source, flow along a 15 cm long by 2 mm diameter flow tube, and expand into a time-of-flight mass spectrometer equipped with laser ionization. Delayed ionization rates for particular clusters are measured as a function of both the photoionization laser wavelength and the temperature of the flow tube. The latter may be varied over the range 77 to 700 K. Klotts has analyzed evaporative dissociation processes on a systematic basis (38) and has presented an analysis to determine statistical behavior (39). In brief, the procedure requires the construction of an "Arrhenius like plot" in which an effective temperature, T , for the microcannonical ensemble is defined. Figure 2 is the result of an analysis on the basis of reference 39 and provides further evidence that the delayed ionization is a statistical process. The ionization rate is plotted against the T which is varied by varying the vibrational temperature of the metal clusters, via the temperature of the flow tube and the wavelength of the ionization laser. As Figure 2 shows, the temperature of the flow tube has a marked influence on the delayed ionization decay rate. This is the first direct evidence for the involvement of the vibrational degrees of freedom in the ionization process and supports the mechanism presented above. intemaI

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch011

143

i n t c m a l

Figure 2. Thermionic Emission Rates for Nb excited at 266 (circles) and 300 nm, (triangles). 9

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

144

LASER CHEMISTRY O F ORGANOMETALLICS

A statistically determined rate constant is of course an excellent "thermometer" for the vibrational degrees of freedom of a molecule. Figure 2 may therefore be used to define the internal vibrational temperature of the niobium clusters prior to photoexcitation. Furthermore, by varying the photoexcitation wavelength, we change T in a precisely determined way. Hence, we may calibrate our "thermometer". When we make the assumption that the vibrational degrees of freedom of the clusters have the same temperature as the walls of the flow tube, both before and after the expansion into the time-of-flight mass spectrometer (as was done in producing Figure 2) we find that the data from different wavelengths lie on the same line. The implication of this observation is that the vibrational degrees of freedom of the metal clusters are not significantly cooled in our weak supersonic expansion. Therefore we have devised a simple and useful method of determining the vibrational temperature of metal clusters on metal cluster beams and have shown that vibrational temperature may be employed as a well defined variable in suitably designed experiments employing supersonic jet sources of metal clusters.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch011

i n t e m a l

Conclusions In this short contribution we have shown that much progress has been made in the use simple metal clusters as model systems for metal-ligand interactions. Futhermore, we have shown that metal clusters sometimes reveal novel molecular science. Much work remains to be done to establish the geometrical and electronic structures for metal clusters, particularly for transition series elements. However, there are already remarkable indications that geometrical structure does play a significant role in the chemistry of these species and that these molecules have a rich and selective chemistry. The challenges for future work in this field are many and varied. Among these are the central problems of the structure of the metal cluster and the structure of the ligand following interaction with the cluster. Progress in the interpretation of chemical reactivity of metal clusters may be aided if the experimental conditions can be controlled so that well characterized chemical measurements can be reported. In this regard, our measurements of the vibrational temperature of niobium cluster beams may be helpful. Acknowledgements This research was supported in part by the Network of Centres Excellence in Molecular and Interfacial Dynamics (CEMAID) in association with the National Research Council of Canada (NRC) and the Natural Sciences and Engineering Research Council of Canada (NSERCC).

Literature Cited 1. Andres, R. P.; Averback, R. S.; Brown, W. L.; Brus, L. E.; Goddard III, W. Α.; Kaldor, Α.; Louie, S. G.; Moscovits, M.; Peercy, P. S.; Riley, S. J.; Siegel, R. W.; Spaepen F.; Wang, Y. J. Mater. Res. 1989, 4, 704. 2. Moscovits, M. Acc. Chem. Res. 1979, 12, 229.

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch011

11.

AMREIN E T AL.

Small Clusters of Refractory Metals

145

3. Whetten, R. L.; Zakin, M.R.; Cox, D. M.; Trevor, D. J.; Kaldor, A. J. Chem. Phys. 1986, 85, 1697. 4. Knickelbein M. B.; Yang, S. J. Chem. Phys. 1990, 93, 5760. 5. Moscovits, M. Ann. Rev. Phys. Chem. 1991, 42, 465. 6. Leblanc, R. P.; Vanbrugghe, B. C.; Girouard, F. E. Can. J. Chem. 1974, 52, 1589. 7. Knickelbein M. B.; Yang, S. J. Chem. Phys. 1990, 93, 1476. 8. Knight, W. D.; Clemenger, K.; deHeer, W. Α.; Saunders, W. Α.; Chou, M. Y.; Cohen, M. L. Phys. Rev. Lett. 1984, 52, 2141. 9. Geusic, M. E.; Morse, M. D.; Smalley, R. E. J. Chem. Phys. 1985, 82, 590. 10. Morse, M. D.; Geusic, M. E.; Heath, J. R.; Smalley, R. E. J. Chem. Phys. 1985, 83, 2293. 11. Cox, D. M.; Reichmann, K.C.;Trevor, D. J.; Kaldor, A. J. Chem. Phys. 1988, 88, 111. 12. Whetten, R. L.; Cox, D. M.; Trevor, D.J.;Kaldor, A. Phys. Rev. Lett. 1985, 54, 1494. 13. Phillips, J. C. J. Chem. Phys. 1986 84, 1951. 14. Kharas, K. C. C. Chem. Phys. Lett. 1989, 161, 339. 15. Zakin, M. R.; Cox, D. M.; Trevor, D. J.; Whetten, R. L.; Kaldor, A. Chem. Phys. Lett. 1987, 133, 223. 16. Hamrick, Y. M.; Taylor, S.; Lemire, G. W.; Fu, Z.-W.; Shui, J.-C.; Morse, M. D. J. Chem. Phys. 1988, 88, 4095. 17. Hamrick Y. M.; Morse, M. D. J. Phys. Chem. 1989, 93, 6494. 18. Zheng, L.-S.; Brucat, P. J.; Pettiette, C. L.; Yang, S.; Smalley, R. E. J. Chem. Phys. 1985, 83, 4273. 19. Elkind, J. L.; Weiss, F. D.; Alford, J. M.; Laaksonen, R. T.; Smalley, R. E. J. Chem. Phys. 1988, 88, 5215. 20. Zakin, M. R.; Brickman, R. O.; Cox, D. M.; Kaldor, A. J. Chem. Phys. 1988, 88, 3555. 21. St. Pierre, R. J.; Chronister, E. L.; El-Sayed, M. A. SPIE 1987, 742, 122. 22. St. Pierre R. J.; El-Sayed, M. A. J. Phys. Chem. 1987, 91, 763. 23. Zakin, M. R.; Cox, D. M.; Kaldor, A. J. Phys. Chem. 1987, 91, 5244. 24. Zakin, M.R.; Brickman, R. O.; Cox, D. M.; Kaldor, A. Chem. Phys. 1988, 88, 5943. 25. St. Pierre, R. J.; Chronister, E. L.; El-Sayed, M. A. J. Phys. Chem. 1987, 91, 5228. 26. Song L.; El-Sayed, M. A. Chem. Phys. Lett. 1988, 152, 281. 27. Brucat, P. J.; Zheng, L.-S.; Pettiette, C. L.; Yang, S.; Smalley, R. E.J.Chem. Phys. 1986,84,3078. 28. Loh, S. K.; Lian, L.; Hales, D.; Armentrout, P. B. J. Chem. Phys. 1988, 89, 3388. 29. Loh, S. K.; Lian, L.; Armentrout, P. B. J. Am. Chem. Soc. 1989, 111, 3167. 30. Loh, S. K.; Lian, L.; Armentrout, P. B. J. Chem. Phys. 1989, 91, 6148. 31. Hales, D. Α.; Lian,L; Armentrout, P. B. Int. J. Mass Spectrom. and Ion Proc. 1990, 102, 269. 32 Miedma, A. R. Faraday Symp. Chem. Soc. 1980, 14, 136. In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

146

LASER CHEMISTRY OF ORGANOMETALLICS

33. Amrein, Α.; Simpson, R.; Hackett, P. A. J. Chem. Phys. 1991, 94, 4663. 34. Amrein, Α.; Simpson, R.; Hackett, P. A. J. Chem. Phys. 1991, 95, 1781. 35. Leisner, T.; Athenassenas, K.; Echt, O.; Kandler, O.; Kreisle, D.; Recknagel, E. Z. Phys .D. 1991, 20, 127. 36. Leisner, T.; Athenassenas, K.; Kandler, O.; Kreisle, D.; Recknagel, E.; Echt, O. Mater. Res. Soc. Proc. Symp. 1991, 206, 259. 37. Wurz, P.; Lykke, K. R. J. Chem. Phys. 1991, 95, 7008. 38. Klotts, C. Ε. Z. Phys. D. 1991, 20, 105. 39. Klotts, C. E. Chem. Phys. Lett. 1991, 186, 73. January 13, 1993

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch011

RECEIVED

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

Chapter 12

Bond Energies and Reaction Kinetics of Coordinatively Unsaturated Metal Carbonyls Eric Weitz, J. R. Wells, R. J. Ryther, and P. House

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch012

Department of Chemistry, Northwestern University, Evanston, IL 60208

Recent measurements of rate constants for reactions of coordinatively unsaturated iron carbonyls with Fe(CO) obtained using time resolved infrared spectroscopy in the gas phase are presented and discussed. These measurements are consistent with the hypothesis that in these and related reactions, the spin states of the reactants and products can significantly influence the magnitude of the rate constant. Measurements of the rates of reaction of Fe(CO) with O and H provide further support for this conclusion. However, variations in rate constants for spin allowed reactions of coordinatively unsaturated metal carbonyls indicate that other factors can significantly influence the magnitude of rate constants for association reactions. Also discussed is a technique for the determination of bond dissociation energies of organometallic compounds. This technique has been used to determine bond enthalpies in systems with weak bonds, such as the Group VI metal carbonyls bound to Xe. Use of time resolved fourier transform infrared (TRFTIR) spectroscopy to probe bond dissociation processes allows for extension of this technique to long time scales and more strongly bound systems. The bond enthalpy for N O bound to W(CO) has been determined using TRFTIR as the kinetic probe. Implications of significant bond enthalpies for typical solvents interacting with coordinatively unsaturated metal carbonyls are discussed in the context of measurements of rate constants for association reactions in solution. 5

4

2

2

2

5

Over the last decade the photochemistry, spectroscopy, and reactivity of metal carbonyls has been the subject of intense interest (1-3). It has been found that metal carbonyls undergo a wide range of facile photochemical reactions. However, it is still the case that, in general, the pathways for these reactions have

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

148

LASER CHEMISTRY OF ORGANOMETALLICS

not been fully characterized. A motivation for study of these systems is that the coordinatively unsaturated metal carbonyls that are the products of many of these reactions can act as efficient catalysts. For example, Fe(CO) can be used to photolytically generate efficient catalysts for olefin isomerization, hydrogénation and hydrosilation reactions (3-6). In many cases, even the initial photoproducts in these reactions have not been characterized. Studies of the product distribution that result from photolytic excitation of organometallic compounds and the iron carbonyls in particular have also been an area of current interest as has been the mechanism of ligand loss following such excitation (1,7-13). Much of the difficulty in characterizing either photoproducts or reaction intermediates arises from their extreme reactivity. For example, it is generally agreed that Cr(CO) coordinates even hydrocarbon solvents on a sub-nanosecond timescale (14,15). Techniques such as matrix isolation spectroscopy have provided considerable information about highly reactive species including coordinatively unsaturated organometallics. Much of what is known about the spectroscopy and structure of these species comes from matrix isolation work (1,2). However, matrix techniques are very limited when kinetic information is desired. Additionally, the possibility of coordination of even relatively inert matrix gases, leading to a change in structure of the coordinated versus uncoordinated species exists. It has long been recognized that it would be desirable to obtain real time kinetic information on coordinatively unsaturated organometallic compounds. Since coordination of hydrocarbon solvents by at least some coordinatively unsaturated organometallic species is a reality (14-16) and data is presented in this work that demonstrates that "solvents" as inert as Kr can be coordinated, the number of solvents that are truly non-coordinating may be quite limited. If a coordinatively unsaturated species has coordinated solvent, putative measurements of association rate constants can be effected by the finite rate of loss of the associated solvent molecules. As such it is desirable to be able to perform spectroscopic and kinetic measurements on "naked" coordinatively unsaturated species. This points to the gas phase as the phase of choice for such measurements. In this environment, these species can be studied in the absence of solvent molecules. However, since coordinatively unsaturated species are very reactive, fast and sensitive spectroscopic techniques are needed for such studies. Since gas kinetic collisions occur on a microsecond timescale at a density of approximately 10 molecules/cc, a system that is capable of detecting considerably fewer than 10 molecules/cc on a submicrosecond timescale is desirable. Transient absorption systems based on a U V or visible probe that meet these criteria have been available for many years. However, many organometallic compounds and metal carbonyls, in particular, are extremely photolabile and their UV-visible spectra are not very structure sensitive (1-3). On the other hand, infrared spectra of the metal carbonyl absorption region is highly structure sensitive (1-3). Thus, time resolved IR is the spectroscopic probe of choice in these systems. Unfortunately infrared detectors are orders of magnitude less sensitive and slower than typical phototubes. Thus, time resolved infrared techniques have historically been plagued by a slow time response and/or a lack of sensitivity. Nevertheless, 5

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch012

5

15

15

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

12.

Coordinatively Unsaturated Metal Carbonyls

WEITZ E T AL.

149

with advances in light sources and infrared detectors it has been possible to develop transient infrared spectrometers that are fast enough and sensitive enough to probe reactions of coordinatively unsaturated organometallic compounds that take place at the gas kinetic limit in a low pressure gas phase environment. Much of the prior work in this area has been well summarized in a number of review articles (1,2). One of the interesting findings that comes out of this work is the uniformly large rate constants for the reaction of most coordinatively unsaturated metal carbonyls with the ligands studied to date. Most of these rate constants are within an order of magnitude of the gas kinetic limit and some are within a factor of two of this limit. A n exception is the rate constant for reaction of Fe(CO) with CO. July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch012

4

Fe(CO) + CO

> Fe(CO)

4

5

This reaction is almost 4 orders of magnitude slower than the gas kinetic limit. The source of this relatively small rate constant has been postulated to be the change in spin state in going from Fe(CO) , with a triplet ground state, to Fe(CO) , with a singlet ground state. Since a similar spin change has not been identified in other systems, the iron pentacarbonyl system currently offers the only available choice for additional study of the effect of this change in spin on rates and pathways for reaction. Thus, we felt it would be interesting to observe and quantify the effect of spin on reactions of coordinatively unsaturated metal carbonyls in the iron system with parent, C O and other small ligands. Much of this work has recently been reported on in detail in other publications (12,13). We also report on a method that allows for the direct determination of bond energies based on measurements of the rate of dissociative loss of a ligand from the system under study. With this information and the rate constants for the association reaction of the relevant coordinatively unsaturated species with the ligand under study and with CO, the bond dissociation energy can be determined. This method has been applied to determine the bond energies for Group V I metal pentacarbonyls species complexed with heavy rare gases and allows for direct measurements of interactions between coordinatively unsaturated organometallic species and a wide variety of ligands (17). A n area of particular interest to us is quantification of the interactions between coordinatively unsaturated metal carbonyls and typical solvent molecules. We have also shown that measurements of bond energies and association rate constants allow us to determine whether a "poor" ligand does not form stable complexes because of kinetic or thermodynamic factors. We have applied this to both N 0 and C F C 1 to show that thermodynamic factors dominate in the making these "poor" ligands for binding to W(CO) (18). For the N 0 system we have also applied time resolved Fourier transform infrared (TRFTIR) spectroscopy to the system which allows for extention of the convenient time range of our laser based systems to longer timescales. 4

5

2

2

2

5

2

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

150

LASER CHEMISTRY O F ORGANOMETALLICS

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch012

Experimental In the experiments under consideration we have used three different experimental set-ups. A l l the systems use either an excimer laser or tripled N d : Y A G laser to provide U V radiation to initiate the photolysis process leading to loss of one or more ligands. Since the wavelength of XeF excimer radiation at 351 nm and that for the frequency tripled output of the N d : Y A G laser at 355 nm are very similar it is not surprising that we have not observed any significant differences in product distributions with these two systems. The excimer laser has also been operated at 248 nm (KrF) and 193 nm (ArF). The N d : Y A G has been used in conjunction with a CO probe laser and the excimer has been used with either a CO or diode laser as the probe. For each of these systems the remainder of the experimental apparatus is similar and has been described in detail in prior publications (1,12,13,17-20). Briefly, the output of either photolysis source makes a single pass through the sample cell. The photolysis laser pulses, at energies of typically less than 10 mj/cm , fill the entire cell volume. Various cell configurations were used with some cells having the capability for use of a purge gas to protect the cell windows from build up of involatile products. The specific advantages and difficulties associated with the use of such a configuration have been discussed in detail in previous publications (12,13). When flow cells were used the replenishment rate of the cell was typically 1-2 Hz with a photolysis rate of 1 Hz. When static cells were used the contents were manually replenished before a significant fraction of the contents of the cell were depleted. Static cells were typically employed when the reaction under study lead to regeneration of close to 100% of the starting material. CO is always a product of photolysis of a metal carbonyl. This CO can be produced in vibrationally and rotationally excited states. Thus CO, along with the photogenerated coordinatively unsaturated metal carbonyls can produce a transient absorption signal. Since the CO laser by necessity operates on transitions that can be absorbed by internally excited CO it is possible for these CO absorptions to obscure the underlying absorptions of photogenerated coordinatively unsaturated metal carbonyls. This can be a significant problem when highly excited CO is produced that overlaps much of the spectral region in which the coordinatively unsaturated metal carbonyls also absorb. This problem was encountered in the Fe(CO) system, particularly when a K r F or ArF laser was used as the photolysis source. This led us to use a diode laser for further study of this system (12,13). With the diode laser, which is continuously tunable, the frequency of operation can be chosen so that it does not overlap the discrete sets of frequencies that correspond to transitions for CO in various (v,J) states. The diode laser can also be tuned to higher frequencies than are convenient or possible with the CO laser. This facilitated the study of absorptions of a number of polynuclear metal carbonyls produced in reactions taking place in the Fe(CO) system. The TRFTIR apparatus is based on a Mattson rapid scan interferometer. In this system the entire interferrogram is recorded for a single laser pulse. We are currently able to acquire a spectrum in 27 msec at 8 cm resolution and this 2

5

5

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch012

12.

Coordinatively Unsaturated Metal Carbonyls

WEITZ E T A L .

151

acquisition procedure can be repeated at 60 msec intervals. There is a reciprocal trade-off between acquisition time and wavelength resolution. We view this system as complementary to our laser based TRIR systems. In the laser based systems long term noise places an effective limit on the timescale over which we can acquire a signal with suitable signal to noise levels. Though this timescale is a function of the wavelength and whether we use the diode laser or C O laser, signals with time constants longer than 10 msec are difficult to acquire without significant low frequency noise. Experiments with the FI1R are implemented in an analogous manner to those using the laser based probes with the FTIR beam replacing the probe laser beam. The FTIR beam is incident on an external HgCdTe detector while the CO and diode laser beams are incident on InSb detectors with time response in the sub 100 nsec range. In the TRIR experiments the probe device monitors the concentration of parent, reactant(s) and product(s) as a function of time. A transient spectrum is produced by having a computer connect points on transient signal taken at different wavelengths that are at a common time delay following the photolysis pulse. The time delay is then incremented to produce a nested set of spectra as shown in Figure 1. In all the figures in this paper, a positive going signal corresponds to less light falling on the detector and thus to production of a new species. A negative going feature corresponds to more light on the detector and thus less absorption by a species present in the cell. For the TRFTIR experiments, manufacturer provided software was used to produce spectra very similar to those described above. However, for this system the interferrogram provided all the data points at a given time delay and subsequent interferrograms were the source of spectra at subsequent times. Thus, the basic difference between the two acquisition modes is that the TRIR system produces signals dense in time and sparse in wavelength (though with the diode laser the individual transients can be acquired at any arbitrary wavelength resolution) and the TRFTIR produces signals that are dense in wavelength and sparse in time. Other manufacturer provided software allows for the determination of decay rates at a given wavelength thus facilitating kinetic measurements. When temperature dependent measurements were carried out, a jacketed cell was used with direct measurement of temperature using a number of thermocouples placed both outside and inside the cell. Temperature reproducibility is within 1°C and accuracy is estimated as ±1°C. Cr(CO) , Mo(CO) , and W(CO) were obtained from Strem Chemicals, Alpha Products, and Aldrich Chemicals, respectively. Fe(CO) was obtained from Alpha Products and all metal carbonyls were sublimed in situ. C O with a stated purity of 99.995% and N 0 and C F C 1 with the stated purities of 99% were obtained from Matheson and used without further purification. 6

6

6

5

2

2

2

Results Reactions of Coordinatively Unsaturated Iron Carbonyls. We have monitored the reactions of coordinatively unsaturated iron carbonyls with parent and CO. Typical results are shown in Figure 1 for the reaction of Fe(CO) with parent. 2

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

152

LASER CHEMISTRY OF ORGANOMETALLICS

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch012

t

2100

2060

2020

1980

1940 -

WaYenumbers (cm

1

3

1900

)

Figure 1. Time-resolved spectra generated upon 193-nm photolysis of Fe(CO) for a sample containing 10 mTorr of Fe(CO) and 10 Torr of argon buffer gas. Panels A - C are segmented into 10 equal time intervals with ranges of (A) 1.0 ps, (B) 10.0 ps, and (C) 100.0 ps. Panel D is segmented into 4 equal time intervals over a range of 200 ps. The ordinate is in arbitrary absorbance units with the zero of absorbance indicated. In panels A and Β the first three traces are labelled. Arrows indicate the direction of change of various features during the indicated time intervals. See text for assignments. (Reprinted with permission from reference 12. Copyright 1991.) 5

5

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

12.

153

Coordinatively Unsaturated Metal Carbonyls

WEITZ ET AL.

The absorptions of the various coordinatively unsaturated iron carbonyls have been identified as a result of previous work (12,13). Products of reactions of coordinatively unsaturated iron carbonyls with parent can be identified since their rate of formation is the same as the rate of loss of parent and reactant. This procedure is illustrated in Figure 2 which displays plots of the rate of loss of Fe(CO) and Fe(CO) and the rate of formation of the reaction product, identified as Fe (CO) . These plots demonstrate that the reaction of Fe(CO) + Fe(CO) to form Fe (CO) takes place. The effect of this reaction can also be seen in Figure 1 where the species labeled as 2 (Fe(CO) ) reacts with parent (5) to form the species labeled as 7 (Fe (CO) ). Assignments of polynuclear species can be further confirmed by observing the reaction of unsaturated polynuclear species with CO to form another polynuclear species which has previously been identified (1). The effect of this type of reaction can also been seen in Figure 1 where feature 7 decays at the same rate that feature 8 appears. Feature 8 has been shown to be due to the unbridged isomer of Fe (CO) . As expected, Features 7 and 8 also form as a consequence of the reaction of Fe(CO) produced following 248 nm photolysis of Fe(CO) . Another example of the latter identification procedure is shown in Figure 3 where the unbridged isomer of Fe (CO) , which forms as a product of the reaction of Fe(CO) with Fe(CO) following 351 nm photolysis of Fe(CO) , is observed to react with CO to form the known stable species, Fe (CO) . The unbridged isomer of Fe (CO) has been identified based on its kinetic behavior and as a result of matrix isolation studies of isomers of Fe (CO) (12). Note the isobestic point in Figure 3 which is characteristic of a direct A —» Β reaction. Using these techniques we have been able to measure the rate constants for reaction of various coordinatively unsaturated iron carbonyls. A summary of these rate constants is shown in Table 1. 2

2

5

7

2

2

5

7

2

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch012

2

7

2

8

2

5

2

3

8

5

5

2

2

9

8

2

8

Measurement of Bond Energies. We have recently implemented a procedure to directly measure the bond dissociation energies of ligands bound to coordinatively unsaturated organometallic species (17,18). The procedure is based on the following kinetic scheme where an appropriate parent molecule is photolyzed in the presence of the ligand under study and CO. hv W(CO)

> W(CO) + CO

(1)

^ ° > W(CO)

6

(2)

W(CO) L

(3)

2-> W(CO) + L

(4)

6

5

W(CO) + CO 5

k

W(CO) + L k

W(CO) L 5

L

=->

5

5

d 5

In equations 1-4 W(CO) is used as a prototypical parent molecule. This scheme involves a competition between CO and L for reaction with W(CO) . C O can react with W(CO) to regenerated W(CO) . When L reacts with W(CO) it forms 6

5

5

6

5

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

154

LASER CHEMISTRY OF ORGANOMETALLICS

0.40

0.35

0.30

0.25

0.20

I ω

0.15

^

ω 0.10

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch012

03

0.05 0.00 Ο

5

10

15

Pressure F e ( C O )

5

2 0

χ 10

2 5

3 0

(Torr)

Figure 2. A plot of the pseudo-first-order rate constant for the reaction of Fe(CO) + Fe(CO) . The rate of decay of Fe(CO) at 1912 cm" (·) and the rate of formation of Fe (CO) at 2044 cm" (Δ) are plotted against the pressure of Fe(CO) . The average slope of the linear least-squares fits gives a rate constant of 12 ± 1.8 ps^Torr" or (3.7 ±0.6) χ 10" cm molecule" V . (Reprinted with permission from reference 12. Copyright 1991.) 1

2

5

2

1

2

7

5

1

10

r

3

1

t \

8u

\

o.o 200 μ,β/interval 2080

2070

2060

I 2050

Wayenumbers (cm

_ 1

2040

)

Figure 3. Time-resolved spectrum generated upon 351-nm photolysis of Fe(CO) . The photolysis flow cell contained 0.5 Torr of Fe(CO) with 5 Torr of CO and 1.25 Torr of argon buffer gas. The data are shown in 10 equal intervals over a range of 2.0 ms. the reaction of unbridged Fe (CO) + C O -> Fe (CO) leads to decay of feature 8u, assigned to unbridged Fe (CO) in the text, and the rise of feature 9, assigned to ground-state Fe (CO) . Arrows indicate the direction of change of the absorption features. (Reprinted with permission from reference 12. Copyright 1991.) 5

5

2

2

9

8

2

2

8

9

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

12.

WEITZ E T A L .

Coordinatively Unsaturated Metal Carbonyls

155

W(CO) L. If L is weakly bound to W(CO) , W(CO) L may dissociate and allow another opportunity for reaction of W(CO) with either L or CO. Since W(CO) is stable on the timescale of these experiments the reaction of W(CO) + CO to form W(CO) is considered to be irreversible (17). Reactant pressures and temperatures are typically chosen so that the rate of dissociation of W(CO) L is rate limiting in the reaction to regenerate W(CO) . Under these circumstances, the kinetic scheme consisting of reactions (1-4) can be solved for the rate of reformation of W(CO) using the steady state approximation. This leads to an expression for the rate of regeneration of W(CO) which we designate as k where 5

5

5

5

6

5

6

5

6

6

6

^

k

•

e

)

l

c

o

' kcotCO] + k [ X e ]

o b s

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch012

f

o b s

m

Xe

This expression can be rearranged to give an expression for 1^, the rate constant for reaction (4). i k

d

"

k

+

obs

k [Xe] Xe

(6)

keotCO]

A measurement of the temperature dependence of k along with a direct measurement of the temperature dependence of the rate constant for the reverse reaction (3) can be used to obtain the temperature dependence of the equilibrium constant for the formation of W(CO) L. This of course can be used to obtain ΔΗ for the W(CO) L bond. In practice, the reaction of coordinatively unsaturated species with either CO or L is typically unactivated indicating a reaction occurring without a barrier. Under these circumstances, the temperature dependence of k can be used to obtain the activation energy for reaction 4 which can be appropriately converted to a bond enthalpy (18). A typical signal for k is shown in Figure 4 where the reaction of W(CO) with Xe is occurring and regeneration of W(CO) is being monitored. From the temperature dependence of k and the rate constants for reaction of W(CO) with Xe and CO it was possible to determine that the bond energy for the W(CO) Xe complex and the other group VI M(CO) Xe complexes is 8.5 ± 0.5 kcal/mole (17). We have also employed this method in an effort to measure the bond dissociation energy for the W(CO) (N 0) complex. N 0 reacts rapidly with W(CO) , but at or near room temperature the rate of dissociative loss of N 0 was too slow to be conveniently followed using a C O laser (18). However, the kinetics of dissociative loss of N 0 from W(CO) (N 0) can be monitored using TRFTIR. A time resolved spectrum showing the disappearance of the absorption due to the W(CO) (N 0) complex is shown in Figure 5. The temperature dependence of this rate constant is shown in Figure 6. From this data, taken in conjunction with previous measurements of the association rate constants for CO and N 0 , a bond dissociation enthalpy of 14.6 ± 0.5 kcal/mole and a d

5

5

d

o b s

5

6

d

5

5

5

5

2

2

5

2

2

5

5

2

2

2

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

156

LASER CHEMISTRY O F ORGANOMETALLICS

Table I. Rate Constants for Reactions of Coordinatively Unsaturated Iron Carbonyls l

Reaction Rates, cm- molecule s +CO +Fe(CO) 3

Species

Fe(CO)

(8.6±2.5) x Κ Γ

Fe(CO)

10

(3.7±0.6) Χ 1 ( Γ (2.9±0.4) x Ι Ο "

10

3

b

(5.2±1.5) x 10"

13

Fe(CO) *

(1.8±0.3) Χ 10"

10

Fe(CO)

4

4

a

yes

(3.0±0.5) x 10" 11

yes

(2.2±0.33)x 10"-11

yes

(5.2±1.2) X 10" 14

no

a

yes

1 1

2

Fe(CO) July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch012

Spin Conserving

1

5

Fe2(CO)

a

(9 ± 3) x 10"

Fe2(CO) (unbridged)

a

(7.8±2.3) x 10~

7

8

yes

11

15

no

a. Not studied. b. Rates of reactions of Fe(CO) with H and 0 have also been measured. = excited state 4

^

2

2

-9. 1

Ο I

0.0

1.0

3.0

5.0

Time (με)

1

Figure 4. Transient signal obtained at 2000 cm" when 5-10 millitorr of W(CO) , 5.2 Torr Xe, 80 Torr He, and 1.2 Torr CO are photolyzed with 355-nm radiation. Decrease is due to loss of W(CO) while increase in signal is return of W(CO) . Line drawn through slow part of W(CO) regeneration is the computed fit, k . (Reprinted with permission from ref. 17. Copyright 1992.) 6

6

6

6

o b s

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch012

12.

WEITZ ET A L .

2060

Coordinatively Unsaturated Metal Carbonyls

2020

1980

1940

157

1900

Wavenumbers

Figure 5. Time-resolved FTIR spectrum of the loss of a W(CO) (N 0) complex (1982 cm ) leading to regeneration of parent (2000 cm ). The photolysis cell contained 76 Torr N 0 and 2 Torr of CO in addition to 10 millitorr of W(CO) . The spectra which are shown at 90 msec time intervals are the result of an average of 200 shots from a frequency tripled Y A G laser with an output at 355 nm. These spectra were acquired at 23 ° C. 5

1

2

1

2

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

6

LASER CHEMISTRY OF ORGANOMETALLICS

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch012

158

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

12.

159

Coordinatively Unsaturated Metal Carbonyls

WEITZ ET AL.

preexponential of 1 χ 10 complex.

12

sec

1

can be determined for the W(CO) (N 0) 5

2

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch012

Discussion Reactions That Produce Polynuclear Iron Carbonyls. Reactions of coordinatively unsaturated metal carbonyls with parent to form polynuclear carbonyls have been studied in a number of systems. In virtually all of these systems, except the iron system, all the measured rate constants are within an order of magnitude of the gas kinetic limit and many are within a factor of two of that limit (1,12). This is also true for coordinatively unsaturated species generated from other members of the Group VIII pentacarbonyls. Studies of reactions of coordinatively unsaturated osmium and ruthenium carbonyls with parent show they are very fast (19,20). The reaction of CO with the tetracarbonyls of the other group VIII metals, which have been calculated to have a C geometry, as does Fe(CO) , are fast (19,20). Thus, the iron carbonyl system appears to be unique. As previously mentioned, this is also the system that, to date, exhibits the smallest rate constant for addition of CO to a mononuclear coordinatively unsaturated metal carbonyl. These observations have been interpreted in terms of the effect of spin conservation on rate constants (1). The ground state of Fe(CO) has been experimentally determined to be a triplet (21). The ground state of Fe(CO) has been calculated to be a triplet (22). Thus addition of C O to Fe(CO) to form Fe(CO) will be spin allowed but the addition of CO to Fe(CO) to form Fe(CO) is spin disallowed. This could explain the source of the almost three orders of magnitude difference in the rate constants for these processes. The rate constant for addition of CO to Fe(CO) is of the same magnitude as the rate constant for addition of C O to Fe(CO) . To explain this it has been postulated that Fe(CO) also has a triplet ground state (1). The rate constants we have measured for the reactions of coordinatively unsaturated metal carbonyls with parent and CO and for C O with polynuclear coordinatively unsaturated carbonyls, are consistent with this hypothesis. For example Fe(CO) reacts with parent at a near gas kinetic rate to form the unbridged isomer of Fe (CO) . Within the context of our picture, this rapid rate of reaction implies spin conservation which in turn implies this isomer of Fe (CO) , first identified in matrix isolation studies, has a triplet ground state (23). Consistent with a triplet ground state, this isomer of Fe (CO) has a relatively small rate constant for reaction with CO to give Fe (CO) , which has a singlet ground state, making this reaction formally spin disallowed. As expected the reaction of Fe(CO) with Fe(CO) to form Fe (CO) is also relatively slow since this reaction is also formally spin disallowed. Excited states of both Fe(CO) and Fe(CO) have been identified in prior studies of the iron pentacarbonyl system (12). The first excited states of these species have been predicted to be singlets (24). Though these species are difficult to study because they are rapidly collisionally relaxed to their ground states, what has been discerned about their reactive behavior indicates that rate constants for reaction are significantly influenced by the spin states of the reactants and 2 v

4

4

3

3

4

4

5

2

3

2

3

2

2

8

8

2

2

4

5

2

3

8

9

9

4

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

160

LASER CHEMISTRY OF ORGANOMETALLICS

products (15). For example, the rate constant for reaction of the excited state of Fe(CO) with parent to form an isomer of Fe (CO) is near gas kinetic and almost three orders of magnitude faster than rate constant for reaction of the ground state of Fe(CO) with parent. As a further test of the influence of spin states of reactants and products on rate constants for reactions, the reaction of H and 0 with Fe(CO) have been measured (12). H was chosen as another small diatomic molecule with a singlet ground state that would provide a complementary system to the Fe(CO) + C O system. 0 was chosen because it has a (Χ Σ~) ground state. Thus a reaction of 0 with Fe(CO) might be expected to have a larger rate constant than a singlet reactant In support of this hypothesis, triplet organic diradicals are typically observed to react rapidly with ground state 0 while analogous singlet diradicals show greatly reduced reactivity (25). The rate of reaction of 0 with Fe(CO) is more than 50 times larger than the rate of reaction of C O with Fe(CO) under the same conditions. Since these studies were intended to compare these reaction rates under similar conditions, they did not involve a complete study of the pressure dependence of the measured rate constant for the reaction with 0 . As the authors point out, since the total pressure in the experiments involving 0 was only 10 Torr it is possible that this reaction was not measured in the high pressure limit (12). If this were the case, the limiting (high pressure) rate constant for the reaction of Fe(CO) with 0 would be even larger and thus would approach the rate for the addition reaction of C O involving spin allowed processes. The rate constant for reaction of H to Fe(CO) to form H Fe(CO) is the same, within experimental error, as that measured for addition of CO to Fe(CO) . This provides evidence that it is not the specific reaction of CO with Fe(CO) that is an anomaly but rather that the class of spin disallowed reactions of Fe(CO) with the small ligands have anomalously small rate constants. The results discussed above are compatible with spin conservation exerting a significant effect on the magnitude of the rate constant for reactions of coordinatively unsaturated metal carbonyls. However, it is clear that this is not the only factor that influences rate constants. There is an order of magnitude difference between rate constants for the formally spin disallowed addition reaction of C O to Fe(CO) and to the unbridged isomer of Fe (CO) (12). Differences in rate constants of similar magnitude have been observed for rapid reactions as well. The reaction of Fe(CO) with ethylene is an order of magnitude faster than the corresponding reaction with CO (26). Reactions of Fe(CO) and Cr(CO) with other olefins are also very fast (27). The detailed explanations for these differences remain to be elucidated. 4

2

9

4

2

2

4

2

4

3

2

2

4

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch012

2

2

4

4

2

2

4

2

4

2

2

4

4

4

4

4

2

8

3

3

4

Bond Dissociation Energies. We have implemented a technique, described in the Results section, that involves a general method for determination of bond dissociation energies of organometallic species. Since laser based probes of reaction kinetics (TRIR) can be used to monitor fast association and/or dissociation reactions the technique can be used to determine bond dissociation energies of even very weakly bound ligands. The use of time resolved fourier transform spectroscopy (TRFTIR) has allowed us to extend our ability to monitor

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

12.

Coordinatively Unsaturated Metal Carbonyls

WEITZ ET AL.

161

dissociative reactions to long time scales leading to a widely applicable technique for determining bond dissociation energies. We have found that even the most weakly interacting of the ligands for which we have been able to measure reaction rates for ligand association reactions have rate constants that are of a similar order of magnitude as rate constants for association reactions involving strongly interacting ligands. It seems likely that it is generally the case that these species are "poor" ligands because thc-y form weak bonds with the coordinatively unsaturated species. However, a bond of over 14 kcal/mole, as has been measured for the bonding of N 0 to W(CO) , is certainly not an insignificant bond energy. In fact this species has a halflife of -.1 sec at 0°C. This is not far from the effective time limit for detection by normal infrared techniques. This bond energy is somewhat less than that estimated in reference 18. It is worthwhile considering why original estimates for the bond dissociation energy of the W(CO) (N 0) complex were higher than the currently reported number. In the original set of experiments, it was not possible to directly measure the rate for dissociative loss of N 0 since it took place on a longer timescale than our laser based system was able to measure but on a shorter time scale than observable with conventional IR spectroscopy. As such our estimate of the rate of regeneration of W(CO) , due to dissociative loss of N 0 , was bounded by these time regimes. Based on these time regimes and the assumption of the same preexponential for dissociative loss of N 0 as we had measured for dissociative loss of another "poor" ligand, CF C1 , we were able to obtain estimates for the activation energy for the loss process and for the corresponding bond enthalpy. Our current measurements directly measure the rate of regeneration of W(CO) and directly determine both the preexponential and the bond enthalpy for the loss of N 0 . Interestingly we find that the preexponential for loss of N 0 from W(CO) is significantly smaller than the preexponential for loss of C F C 1 from the same parent. Though this change in preexponential is likely to reflect information about bonding geometries and force constants, more work is necessary before firm conclusions can be drawn regarding the significance of the change in magnitude of the preexponentials. Data on the binding of Xe and Kr to the Group V I metal pentacarbonyls illustrates the extreme reactivity of coordinatively unsaturated metal carbonyls (17). This leads to the expectation that coordinatively unsaturated organometallic compounds will interact, in some cases strongly, with many solvents. A concern under these circumstances is that the species being studied in solution are not the "naked" coordinatively unsaturated species but a coordinated molecule. Attempts to study association reactions under these circumstances could lead to kinetic data that was influenced by dissociative or associative displacement of the solvent molecule(s). It is interesting to note that even though it may not have been generally recognized that coordinatively unsaturated metal carbonyls are sufficiently reactive to coordinate common solvents, data to support this point of view has been available for over 10 years. In the early '80s Bonneau and Kelly observed a 3 order of magnitude decrease in the rate of addition of C O to Cr(CO) in cyclohexane versus perfluorocycloheptane (16). In 1982 Vaida and Peters reported that Cr(CO) associated with simple solvents on a psec timescale (14).

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch012

2

5

5

2

2

6

2

2

2

2

6

2

2

5

2

2

6

5

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

162

LASER CHEMISTRY O F ORGANOMETALLICS

Part of our future work in this area will involve measurements of the bond dissociation energies of common solvents coordinated to Group VI pentacarbonyls in an effort to quantify how significantly these interactions can and have effected solution phase kinetic studies of these species. We are also interested in further elucidating the bonding mechanism for such species. Summary Rate constants for reactions of coordinatively unsaturated iron carbonyls with Fe(CO) are consistent with conservation of spin playing a role in determining the magnitude of these rate constants. Rates of reaction of H and 0 with Fe(CO) are also consistent with spin conservation being a factor in determining rate constants for reactions of metal carbonyls. Variations in rate constants for spin allowed reactions of metal carbonyls indicate the factors other than spin conservation can significantly effect rate constants for association reactions. A technique to measure bond enthalpies, which involves a real time monitor of the rate of dissociation of a organometallic species, is described. Using this technique measurements have been made of the bond enthalpies for Xe bound to the Group V I pentacarbonyls. Bond enthalpies are 8.5 ± 0.5 kcal/mole for the three pentacarbonyl complexes. Using this technique and time resolved fourier transform infrared spectroscopy, the bond dissociation enthalpy of a W(CO) (N 0) complex has been determined to be 14.1 ± 0.5 kcal/mole. 5

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch012

2

2

4

5

2

Acknowledgments. We acknowledge support of this work by the National Science Foundation under grants, CHE-88-06020 and 90-24509. We also would like to thank the donors of the Petroleum Research Fund administered by the American Chemical Society for partial support of this work under grant 18303AC6,3. We thank Professor Martyn Poliakoff for a number of useful discussions and suggestions which were facilitated by N A T O support under a collaborative grant. We also acknowledge Dr. Paula Bogdan for her role in implementing the technique for determining bond dissociation enthalpies.

References 1. Weitz, E. J. Phys. Chem. 1987, 91, 3945. 2. Poliakoff, M.; Weitz, E. Adv. Organometallic Chem. 1986, 25, 277; Poliakoff, M.; Weitz, E. Acc. Chem. Res. 1987, 20, 408. 3. Geoffroy, G. L.; Wrighton, M. S. Organometallic Photochemistry, Academic NY, 1979. 4. Wrighton, M. S.; Ginley, D. S.; Schroeder, M. Α.; Morse, D. L. Pure Appl. Chem. 1975, 41, 671; Wrighton, M. S.; Ginley, D. S.; Schroeder, M. Α.; Morse, D. L. Pure Appl. Chem. 1975, 41, 671; Whetten, R. L.; Fu, K. J.; Grant, E. R. J. Am. Chem. Soc. 1982, 104, 4270; Weiller, B. H.; Grant, E. R. J. Am. Chem. Soc. 1987, 109, 1051.

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

12.

WEITZ ET AL.

5.

6.

7.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch012

8. 9. 10. 11. 12. 13. 14. 15.

16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

Coordinatively Unsaturated Metal Carbonyls

163

Whetten, R. L.; Fu, K. J.; Grant, E. R.J.Chem. Phys. 1982, 77, 3769;Wrighton, M. S.; Ginley, D. S.; Schroeder, M. Α.; Morse, D. L. Pure Appl. Chem. 1975, 41, 671; Schroeder, M. Α.; Wrighton, M. S. J.Am. Chem. Soc. 1976, 98,551; Schroeder, M. Α.; Wrighton, M. S.J.Am. Chem. Soc. 1976, 98, 551. Mitchener, J. C.; Wrighton, M. S. J. Am. Chem. Soc. 1981, 103, 975; Trusheim, M. R.; Jackson, R. L. J. Phys. Chem. 1983, 87, 1910; Jackson, R. L.; Trusheim, M. R. J. Am. Chem. Soc. 1982, 104, 6590; Darsillo, M. S.; Gafney, H. D.; Paquette, M. S. J. Am. Chem. Soc. 1987, 109, 3275. Ray, U.; Brandow, S. L.; Bandukwalla, G.; Venkataraman, Β. K.; Zhang, Z.; Vernon, M. J. Chem. Phys. 1988, 89, 4092; Venkataraman, Β. K.; Bandukwalla, G.; Zhang, Z.; Vernon, M. J. Chem. Phys. 1989, 90, 5510. Waller, I. M.; Hepburn, J. W. J. Chem. Phys. 1988, 88, 6658; Waller, I. M.; Davis, H. F.; Hepburn, J. W. J. Phys Chem. 1987, 91, 506. Fletcher, T. R.; Rosenfeld, R. N. J. Am. Chem. Soc. 1988, 110, 2097; Rosenfeld, R. N.; Ganske, J. A. J. Phys. Chem. 1989, 93, 1959. Yardley, J. T.; Gitlin, B.; Nathanson, G.; Rosan, A. M. J. Chem. Phys. 1981, 74, 370. Rayner, D. M.; Ishikawa, Y.; Brown, C. E.; Hackett, P. A. J. Chem. Phys. 1991, 94, 5471; Ishikawa, Y.; Brown, C. E.; Hackett, P. Α.; Rayner, D. M. J. Phys. Chem. 1990, 94, 2404. Ryther, R. J.; Weitz, E. J. Phys. Chem. 1991, 95, 9841. Ryther, R. J.; Weitz, E. J. Phys. Chem., 1992, 96, 2561. Welch, J. Α.; Peters, K. S.; Vaida, V. J. Phys. Chem. 1982, 86, 1941. Wang, L.; Zhu, X.; Spears, K. G. J. Am. Chem. Soc. 1988, 110, 8695; Simon, J. D.; Xie, X. J. Phys. Chem. 1986, 90, 6751; Lee, M.; Harris, C. B. J. Am. Chem. Soc. 1989, 111, 8963; Yu, S.; Xu, X.; Lingle, R. Jr.; Hopkins, J. B. J. Am. Chem. Soc. 1990, 112, 3668. Bonneau, R.; Kelley, J. M. J. Am. Chem. Soc. 1980, 102, 1220; Kelly, J. M.; Long, C.; Bonneau, R. J. Phys. Chem. 1983, 87, 3344. Wells, J. R.; Weitz, E. J. Am. Chem. Soc., 1992, 114, 2783. Bogdan, P. L.; Wells, J. R.; Weitz, E., J. Am. Chem. Soc. 1991, 113, 1294. Bogdan; P. L.; Weitz, E. J. Am. Chem. Soc. 1990, 112, 639. Bogdan; P. L.; Weitz, E. J. Am. Chem. Soc. 1990, 111, 3163. Barton, T. J.; Grinter, R.; Thomson, A. J.; Davies, B.; Poliakoff, M.J.Chem. Soc. Chem. Commun. 1977, 841. Daniel, C.; Bénard, M.; Dedieu, Α.; Wiest, R.; Veillard, A J. Phys. Chem. 1984, 88, 4805. Poliakoff, M.; Turner, J. J. J. Chem. Soc. A. 1971, 2403; Fletcher, S. C.; Poliakoff, M.; Turner, J. J. Inorg. Chem. 1986, 25, 3597. Daniel, C.; Bénard, M.; Dedieu, Α.; Wiest, R.; Veillard, A J. Phys. Chem. 1984, 88, 4805. Berson, J. A. In Diradicals; Borden, T. W., Ed.; Wiley-Interscience: New York, 1982; Chapter 4. Hayes, D.; Weitz, E. J. Phys. Chem. 1991, 95, 2723. Gravelle, S. J.; Weitz, E. J. Am. Chem. Soc. 1990, 112, 7839.

RECEIVED

March 30, 1993

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

Chapter 13

Time-Resolved IR Spectroscopy of Transient Organometallic Complexes in Liquid Rare-Gas Solvents Bruce H . Weiller

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch013

Mechanics and Materials Technology Center, The Aerospace Corporation, P.O. Box 92957, Los Angeles, C A 90009

Photolysis of M(CO) (M = Cr and W) in low-temperature, rare-gas solutions leads to the formation of transient organometallic complexes between M(CO) and weak ligands, including CO , N O, Xe and Kr. Time-resolved IR spectroscopy was used to capture IR spectra of the complexes over the temperature range of 150 to 200 K. A detailed kinetic investigation of the reaction of W(CO) Xe with CO in liquid Xe is presented. From the temperature dependence of the kinetics, we obtain the binding enthalpy of Xe to W(CO) : ΔΗ = 8.6 ± 0.4 kcal/mol, in good agreement with independent gas-phase results. This establishes the utility of the technique for determining binding energies of weak ligands to electron-deficient metal centers. 6

5

2

2

5

5

Organometallic chemical vapor deposition (OMCVD) is a widely used technique for the deposition of thin-film materials for electronic, optical, and protective coatings. Central to this process are chemical reactions that occur between the organometallic precursor compounds, added reagents, and the heated surface. Reactions in the gasphase and on the surface can have a profound influence on the properties of the resulting materials. In order to control and modify the properties of the materials produced by OMCVD, it is critical to elucidate the reaction mechanisms involved in the conversion of organometallic precursors to thin-film materials. This includes the structures of the reactive intermediates involved and the rates of their elementary reactions. One reaction common to many OMCVD systems is complex formation between added reagents and electron-deficient metal centers. For example, in the deposition of III-V and II-VI semiconductors such as GaAs and ZnS, the reactants are often Lewis acid-base pairs such as Ga(CH3)3 and ASH3 or Zn(CH3)2 and H2S. Clearly one of the first reactions to occur in these systems will be complex formation. Laser-assisted OMCVD is another area where complex formation is important. Lasers can be used to deposit thin films at low temperatures with spatial selectivity by photolyzing organometallic precursors. Often metal carbonyls are used alone or in mixtures with other reagents in the laser-assisted OMCVD of metals and 0097-6156/93/0530-0164$06.00/0 © 1993 American Chemical Society In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch013

13.

WEILLER

Transient Organometallic Complexes

165

ceramics. For example, Cr02, a magnetic material commonly used in magnetic storage media, can be formed at room temperature by photolysis of mixtures of Cr(CO)6 and O2 or other oxidants (1). Photodissociation of C O from metal carbonyls is facile, and complex formation between coordinatively unsaturated fragments and added reagents is likely. The stability of the complexes formed and the subsequent reaction chemistry will play a significant role in determining the properties of the resulting material. In order to study the intermediates important in O M C V D processes and their chemical reactions, we use a novel approach, laser photolysis in liquid rare-gas sol­ vents with time-resolved IR spectroscopy. This is a powerful combination of tech­ niques that slows reaction rates and provides accurate kinetic parameters by the use of the low temperatures. Furthermore, rare-gas solvents are inert, and the IR spectra of dissolved organometallics are simple with relatively narrow bands and no solvent absorptions. The utility of liquid rare-gas solvents for organometallic reaction studies (2) and for the synthesis of novel organometallic compounds (3) has been demonstrated. The advantage of combining liquid rare-gas solvents with timeresolved IR spectroscopy was shown in recent studies that reported the first direct rate measurement of organometallic C-H bond activation (4). In the studies presented here, we have used a commercial FTIR spectrometer and liquid rare-gas solvents to examine the photochemistry relevant to the formation of metal-oxide films from metal carbonyls and N2O or CO2. During laser photol­ ysis of mixtures of Cr(CO)6 or W(CO)6 d N2O in liquid rare-gas solutions, relatively stable complexes are formed that are implicated in the laser-assisted O M C V D of metal oxides. CO2, which is isoelectronic with N2O, forms a much shorter-lived complex. In addition, we have investigated the complexes formed with the rare gases Xe and Kr. We present the first spectral data for Cr(CO)5Kr and W(CO)5Kr in fluid solution and also a detailed kinetic study for the reaction of W(CO)5Xe with CO. From the temperature-dependent kinetics, we obtain a value for the binding energy of Xe to W(CO)5: 8.6 ± 0.4 kcal/mol (± σ). This result is in excellent agreement with a recent independent gas-phase measurement (5) and demonstrates the utility of the technique for determining binding energies. a n

Experimental In these experiments, we used an excimer laser (Questek) to photolyze metal carbonyls dissolved in liquid rare-gas solvents and a commercial FTIR spectrometer (Nicolet, Model 800) that can be operated in a rapid-scan mode. To liquefy the rare gases, we used a specially designed high-pressure, low-temperature cell, shown in Figure 1. The cell, which was mounted in the sample compartment of the FTIR spectrometer, consists of a copper block with two perpendicular optical axes 1.43 cm and 5.0 cm in length. The cell is enclosed in an evacuated dewar and is cooled with liquid nitrogen. The temperature is measured with silicon diodes (± 0.1 K ) and controlled via heaters. Samples containing metal carbonyls, reactants (N2O, CO2, or CO) and rare gas were prepared on a stainless-steel vacuum line. For the rapid-scan FTIR spectra, the interferometer mirror was moved at high velocity, and the laser was synchronized with the take-data signal of the spectrometer. No signal averaging was used. The laser beam was aligned along the long axis of the cell, and the IR beam passed through the short axis. A shutter (Uniblitz) served to block the laser until the start command was given at the keyboard of the computer. A T T L pulse from the computer then opened the shutter to let a single laser pulse irradiate the sample. Interferograms were then collected in only one mirror direction at equal

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

LASER CHEMISTRY OF ORGANOMETALLICS

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch013

166

Figure 1. Diagram of the high-pressure, low-temperature cell used for IR spectroscopy of rare-gas solutions. Not shown is the external vacuum dewar.

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

13. WEILLER

Transient Organometallic Complexes

167

time intervals after the laser pulse for the desired time period. For the experiments reported herein, spectra were collected at 8 c m resolution and a time interval of 0.09 s using an M C T detector. Spectra were collected in the forward direction only, and the laser was synchronized to fire 2 ms prior to the start of data collection. After data collection, the interferograms were transformed, and converted to absorbance, and the difference was taken with reference spectra. For kinetic measurements, peak areas were used in order to avoid complications from potential distortions in peak widths. C O concentrations were determined from the integrated absorbance of the CO band using the gas-phase band strength (6) corrected for the index of refraction (7). Chemicals were obtained from the following suppliers: Cr(CO)6 and W(CO)6 (Alfa), Xe and K r (Spectra Gases, research grade), CO2 and CO (Matheson, research grade), and N2O (Puritan-Bennett, medical grade). C O was purified with a liquidN2 cooled trap; the other chemicals were used without further purification. July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch013

-1

Results and Discussion M(CO)s(N20) and M(C0)s(C02). In order to understand the deposition of metal oxides by laser-assisted O M C V D , we have started an investigation into the complexes formed between metal carbonyl fragments and N2O and CO2. As discussed in the introduction, metal oxides can be formed at room temperature by photolysis of metal carbonyls and an oxidizer such as O2, N2O or CO2. Relatively little is known about this process, but earlier studies in low-temperature matrices (8) have shown the existence of complexes between M(CO)5 and N2O and CO2. Timeresolved experiments in the gas phase have also confirmed the existence of a complex with N2O (9). M(CO)5(N20). Figure 2 shows the IR spectra obtained when a solution of W(CO)5 and N2O dissolved in liquid K r is photolyzed at 351 nm. The spectrum is presented as the difference between spectra before photolysis and at ~l-min. intervals after photolysis. A new species is formed as shown by the positive peaks at 2092, 1968, and 1954 cm" that slowly decays to reform W(CO)6. Using the rapid-scan capability of the spectrometer, we have also observed N2O complex formation in the gas phase. Very similar results were obtained for Cr(CO)6 and N2O in the gas phase and in solution (Table I). It should be noted that our observation of the high-frequency ai band is the first reported for these complexes and confirms their proposed C4 symmetry. As shown in Table I, our results in the gas phase and in liquid K r solution are in good agreement with independent gas-phase measurements. However, both sets of results diverge significantly from matrix isolation studies. M(CO)5(C02). CO2 is isoelectronic with N2O and can also be used as an oxidant to form metal oxides. When solutions of W(CO)5 and CO2 in liquid K r were photolyzed at 161 K , no stable complexes could be observed. However, using the rapid-scan mode of the spectrometer, it was possible to observe the formation of W(CO)5(C02), as shown in Figure 3. The time between spectra is 0.09 s, and the complex decays orders of magnitude faster than the analogous N2O complex. A reasonable explanation for this observation is a weaker interaction between W(CO)s and CO2. However, a definite explanation will have to wait for planned kinetic studies of the decay process. We can rule out polynulcear species since we see no CO2 complex are in excellent agreement with the matrix isolation data, unlike the case for N2O. The apparent difference in the IR band positions of M(CO)5(N20) in the gasphase and in liquid K r versus those observed in low-temperature matrices is 1

V

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch013

168

LASER CHEMISTRY O F ORGANOMETALLICS

1 2080

ι 2100

1 2060

1 1 1 2040 2020 2000 Wavenumbers

1 1980

1 1960

r 1940

Figure 2. IR spectra of W(CO) (N 0) formed in liquid K r at 180 K . 5

2

Table I. Observed IR Frequencies for CO2 and N2O Complexes Complex

Frequencies (cm- )

W(CO) (N 0)

1977 1980 1968

5

2

2092

Cr(CO) (N 0) 5

2

2089

W(CO) (C0 ) 5

2

Cr(CO) (C0 ) 5

Conditions

Reference

1967 1954

gas, 298 Κ gas, 298 Κ 1. Kr, 180 Κ

this work 9 this work

1950

1925

s. Ar, 10 Κ

8

1979 1972 1950

— 1955 1925

gas, 298 Κ l . K r , 180 Κ s. Ar, 10 Κ

this work this work

1957

1930

l . K r , 180 Κ

this work

1956

1926

s. Ar, 10 K

8

1952

1925

s. Ar, 10 K

8

1

2

--

8

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

13.

WEILLER

Transient Organometallic Complexes

169

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch013

20

ι 2000

ι

ι — 1960 1940 Wavenumbers Figure 3. Time-resolved IR spectra of W(CO)5(C02) in liquid Kr. The times after the laser pulse are shown in the legend. 1980

significant. The data for the CO2 complex and for Xe complexes (see below) are in good agreement with matrix data and eliminates any doubt about the experimental evidence for bridging CO bands. Finally, we note that the observed bands for the technique. Therefore, we must look to other reasons to explain the large difference with the matrix isolation spectra (Δν =18 cm ). One possibility is the formation of different isomers in the higher-temperature experiments (liquid K r or the gas phase) versus the cryogenic ones (matrices). The nature of the bonding in the N2O complexes is an open question, but calculations indicate that the N-bonded isomer is more stable (10). It may be that a less-stable isomer, such as an O-bonded species, is formed in the ultra-cold, rigid matrix that is not stable at higher temperatures (>150 K ) . Further investigations are underway to address this question and to quantify the bond energies in these complexes. -1

M ( C O ) s K r and M(CO)sXe. The question of complex formation between M(CO)5 and rare-gas atoms is not new. Early matrix isolation studies showed, from shifts in the lowest electronic transition, that M(CO)s interacts significantly with Ar, Kr, and Xe (11). The strength of the interaction was correlated with the polarizability of the rare-gas atoms. Direct evidence for a bonding interaction between Mn(CO)s and K r was obtained from EPR spectra in K r matrices (12). Finally, Cr(CO)5Xe was observed in liquid-Xe solutions with an FTIR spectrometer (13). These workers presented a rough estimate of the bond energy for Xe to Cr(CO)5, but, until this symposium, there were no reliable values for the binding energy of Xe to metal centers. Now, from the gas-phase work (5) and this work in liquid Xe, we have two independent measurements of the value for W(CO)5Xe. M(CO)5Kr. In order to determine the role of the solvent in liquid-Kr studies, we have examined the transients formed when W(CO)6 and Cr(CO)6 are photolyzed in liquid K r with no added reagents. During the time-resolved experiments with

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch013

170

LASER CHEMISTRY O F ORGANOMETALLICS

CO2, we observed a short-lived transient prior to the formation of the CO2 complex that was not observed in experiments with large concentrations of CO2. When W(CO)6 was photolyzed in liquid K r in the absence of added reagents, we observed an identical transient, as shown in Figure 4. This new species, which we assign to M(CO)5Kr, is short lived even at 154 K, with a lifetime of ~ 100 ms. Similar results were obtained with Cr(CO)6 and with two samples of K r from different suppliers. Unfortunately, due to the limited solubility of metal carbonyls in liquid Ar, it was not possible to use a more inert solvent than K r to confirm the assignment. The observed frequencies for Cr(CO)5Kr are in agreement with data from matrix isolation experiments and fit the expected trend with polarizability of the rare-gas atom as shown in Table II. A l l of these data are consistent with an assignment of the transients to M(CO)5Kr. To our knowledge, this is the first observation of M(CO)5Kr in fluid solution. M(CO)5Xe. In contrast to the short-lived complexes observed with Kr, com­ plexes with significant stability were formed with Xe. When a solution of Cr(CO)6 in liquid Xe was photolyzed, we observed the formation of a persistent complex that decayed on the time scale of minutes at 180 K. The observed frequencies, shown in Table II, are in excellent agreement with the results of Simpson et al. who demon­ strated that the complex is Cr(CO)sXe (13). Our observation of all three expected IR bands confirms their assignment. The same experiment with W(CO)6 gave a complex with a lifetime of minutes at 198 Κ and three IR bands (Table II). Figure 5 shows that when the analogous experiment was performed using a mixture of 5% Xe in Kr, we observed very similar IR bands and lifetime. On this basis, we assign the complex to W(CO)5Xe. C O Substitution K i n e t i c s in L i q u i d X e . In order to determine the binding energy of X e to W(CO)5, we have examined the kinetics of the reaction of W(CO)5Xe with CO. Figure 6 shows the rapid-scan spectrum when W(CO)6 is photolyzed in a mixture of 0.011 M CO in liquid Xe at 198.0 K. The decay of 10

ι 2050

1 2000

1 1950 Wavenumbers

1 1900

r 1850

Figure 4. Time-resolved IR spectrum of W(CO)sKr in liquid K r at 154 K . In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

13.

WEILLER

Transient Organometallic Complexes

171

Table II. Observed IR Frequencies for M(C0)5(Q) Complexes Complex

Frequencies (cm - ) 1

W(CO) (Ar) 5

W(CO) (Rr) W(CO) (Xe) 5

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch013

5

2090

Cr(CO) (Ar) Cr(CO) (Kr) 5

5

Cr(CO) (Xe) 5

a

2087

Conditions

Reference 8 this work

1969 1962

1935 1933

s. Ar, 10 Κ l . K r , 150 Κ

1958

1930

1. Xe, 170 Κ

this work

1963

1936

1. Kr, 183 Κ

this work

1973

1936

s. Ar, 10 Κ

8

1965 1961

1938

a

1933

l . K r , 180 Κ s. Ar, 10 Κ

this work 11

1960

1934

l . X e , 163 Κ

this work

1960.3

1934

l . X e , 175 Κ

1929

s. Xe, 20 K

13 11

1956 Estimated frequencies from ref. 11.

a

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch013

172

LASER CHEMISTRY O F ORGANOMETALLICS

ι

ι

ι

2020

2000

1980

1 1960

1 1940

1 1920

Γ

1900

Wavenumbers Figure 6. Time-resolved IR spectrum of W(CO)5Xe in liquid Xe with added CO. The time between spectra is 0.09 s. W(CO)5Xe is well represented by an exponential with a decay constant k ^ s = 7.4 ± 0.3 s" , and W(CO>6 recovers at the same rate within error limits. Figures 7 and 8 show the C O concentration dependence of the observed decay constants as a function of temperature from 173.0 Κ to 198.0 K . The decay constants are linear over the concentration range studied (0.003 to 0.03 M) and display roughly a factor of 2 increase every 5 K . At 8 cm" resolution, the time resolution of the spectrometer is -0.1 s, and the fastest rate we can measure is 10 s . The outlying data point in the 198.0 Κ data shows this limit and is not included in the fit. A likely mechanism for this reaction is dissociative substitution in which Xe dissociates from W(CO)5Xe to form W(CO)5, followed by reaction with CO to give W(CO) : 1

1

_1

6

W(CO) Xe — W ( C O ) 5

W(CO) + Xe

5

+ Xe

(1)

> W(CO) Xe

5

(2)

5

W(CO) + CO — W ( C O ) 5

(3)

6

Using the steady-state approximation on the intermediate, W(CO)5, results in an ex­ pression for the observed decay constant (kobs): _ kik [CO] K!k [CO] " k.i[Xe]+k [CO] ~ [Xe] 2

^

2

ά w

2

Here we have assumed that k_i[Xe] » k [CO] because the values for k_i and k are comparable (5b), (14), and [Xe] » [CO]. 2

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

2

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch013

WEILLER

Transient Organometallic Complexes

Figure 7. CO dependence of k o at 198 Κ ( Ο ) , 193 Κ (M ), and 188 Κ (•). The highest [CO] point at 198 Κ is not included in the fit (see text). bs

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

173

174

LASER CHEMISTRY OF ORGANOMETALLICS

Another mechanism that needs to be considered is the associative pathway consisting of a bimolecular displacement of Xe by CO: W(CO)5Xe + CO — W ( C O ) 6 + Xe,

(5)

where k is the bimolecular rate constant. These two mechanisms can often be distinguished from the dependence of k ^ s on C O concentration. The dissociative pathway should show limiting behavior in the concentration of CO. This can be seen from equation (3), which reduces to kobs ~ k l when k2[CO] » k_i[Xe]. For the associative mechanism, kobs = k [CO] * should show linear dependence on [CO]. Since the leaving ligand (Xe) is the solvent in this case, it was not possible with the current apparatus to obtain data under conditions where k2[CO] » k_i[Xe], due to the time-resolution limits of the spectrometer. Therefore, the two mechanisms are kinetically indistinguishable, and we must use other data to identify the mechanism. The pre-exponential factor can be used to distinguish between these two mechanisms. If an associative mechanism is operative, then kobs/[CO] would be equal to the bimolecular rate constant. In this case, we would expect the pre-exponential factor derived from a simple Arrhenius plot to be in the range of 10$ to 10 M l s . Indeed, for Cp*Rh(CO)Xe, a compound expected to undergo associative substitution (15), the measured pre-exponential factor for C O substitution in liquid Xe is log(A) = 8.8 ± 0.3 (16). The value we derive in this work from a plot of ln(kobs/[CO]) vs 1/T, log(A) = 12.2 ± 0.2, is not consistent with an associative mechanism. a

a n c

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch013

a

9

_

- 1

The temperature dependence of the quantity K i k gives the binding energy of Xe to W(CO)5. From the slopes of Figures 7 and 8 and the Xe density, we obtain values for K i k (17). Figure 9 shows that when the quantity ln{Kik } is plotted 2

2

2

Temperature ( K ) 200

195

1900

1180

185

175

170

10

6 5.0

5.2

5.4

5.6 3

5.8

1

T'^xlO" K' ) Figure 9. Arrhenius plot of l n { K i k ) versus 1/T. 2

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

13. WEILLER

175

Transient Organometallic Complexes

versus 1/T, a straight line is observed. The significance of the slope and intercept of the fit can be seen from ln{*TAl = - f - +

ln^)

+

-

(

A

H

(6)

The slope is equal to -(ΔΗι + E2)/R, where ΔΗι is the enthalpy for reaction (1), and E2 is the activation energy for reaction (2). The intercept is equal to {ASi/R + ln(A2)}, where AS\ is the entropy for reaction (1), and A2 is the pre-exponential factor for the rate constant for reaction (2). The elementary rate constant for reaction (2) has been determined to be k2 = 3.0 χ 1 0 M ^ s at 300 Κ in the gas phase (14). Since k2 is within an order of magnitude of the gas kinetic value, it is reasonable to assume that the activation energy for this reaction is not significant (E2 ~ 0). This allows us to determine the binding enthalpy of Xe to W(CO)s; ΔΗι = 8.6 ± 0.4 kcal/mol. In addition, we can use the value for k2 as a measure of A2 in order to determine A S i . While there is some question concerning the relationship between Α-factors in solution and in the gas-phase, a reasonable estimate for A2 is the diffusion-controlled rate constant of 5 χ 10 M ' V (18). This leads to a value of A S i = + 18 cal/molK that is slightly low for dissociation of a heavy atom like Xe. However, this value is not inconsistent with the small bond energy we derive for XeW(CO)5. A weak Xe-W bond implies a low vibrational frequency and a significant amount of associated vibrational entropy that would be lost upon dissociation of Xe. The pre-exponential factor derived from Figure 9 is in good agreement with expectations for the dissociative mechanism. If we use the approximation that k2 and k_i are roughly equal in magnitude, then Kik2 ~ k i . In this case, the intercept of Figure 9 should give a value consistent with a unimolecular Α-factor for k i . The value we find, log(A) = 13.6, is in agreement with this expectation and is inconsistent with a bimolecular Α-factor as discussed above. This further confirms the dissociative substitution mechanism. Our value of ΔΗι = 8.6 ± 0.4 kcal/mol is in excellent agreement with an inde­ pendent determination in the gas phase of 8.5 ± 0.5 kcal/mol that was first presented at this symposium (5a). It should be noted that our approach directly gives a value for ΔΗ1 and does not rely on any assumptions about the activation energy of the reverse reaction. However, we do require information about the activation energy for the CO recombination step. In all cases measured to date except one, Fe(CO)4 + CO, these reactions are so fast that the activation energies are insignificant (19). Since this is one of the first measurements of the X e binding energy to an organometallic complex, it should be evaluated in light of the available related data. First, is a calculation of the binding energy for Mo(CO)sKr of 5.3 kcal/mol based on dispersion forces (20). Since Xe is more polarizable than Kr, we should expect a larger value. Second is thermal desorption data of Xe from metal surfaces; the binding energies of Xe to Pt(l 11) and W ( l 11) have been determined to be 9.2 ± 0.9 (21) and 9.3 ± 1.3 (22) kcal/mol, respectively. Given this range of values, 8.6 ± 0.4 kcal/mol for W(CO)sXe appears to be quite reasonable. 10

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch013

^ *>>

9

- 1

1

Conclusions We have shown that time-resolved IR spectroscopy in liquid rare-gas solvents is a useful technique for studying a range of transient complexes of M(CO)5 with weak ligands, such as N2O, CO2, Xe and Kr. The N2O and CO2 complexes are implicated as intermediates in the formation of metal-oxide thin films by laser-assisted

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

176

LASER CHEMISTRY OF ORGANOMETALLICS

O M C V D . We have presented the IR spectra of Cr(CO)5Kr, W(CO)sKr, and W(CO)5(C02) in fluid solution for the first time. In addition, we have carried out a detailed kinetic investigation of the reaction of W(CO)5Xe with CO in liquid Xe. From the temperature dependence of the kinetics, we obtained the binding enthalpy of Xe to W(CO)5: ΔΗ = 8.6 ± 0.4 kcal/mol. This value is in good agreement with an independent gas-phase determination and establishes the utility of the approach for determining binding energies. Future work will include bond energy measurements for N 0 and C 0 complexes and investigations into the intermediates in the photo­ chemical deposition of metal oxides from metal carbonyls. 2

2

Acknowledgments

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch013

This research was supported by the Aerospace Sponsored Research (ASR) Program.

Literature Cited (1) Perkins, F. K.; Hwang, C.; Onellion, M.; Kim, Y-G.; Dowben, P. Α., Thin Solid Films, 1991,198,317. (2) see for example: Turner, J. J.; Simpson, M. B.; Poliakoff, M.; Maier, W. B., J. Am. Chem. Soc. 1983,105,3898. (3) Sponsler, M. B; Weiller, Β. H.; Stoutland, P. O.; Bergman, R. G.,J.Am Chem. Soc. 1989,111,6841. (4) Weiller, B. H.; Wasserman, E. P.; Bergman, R. G.; Moore, C. B.; Pimentel, G. C.,J.Am. Chem. Soc. 1989,111,8288. (5) a) Weitz, E.; Wells, J. R.; Ryther, J. R.; House, P., this volume, b) Wells, J. R.; Weitz, E., J. Am. Chem. Soc. 1992,114,2783. (6) Pugh, L. Α.; Rao, Κ. N., "Intensities from Infrared Spectra" Molecular Spectroscopy: Modern Research; Academic: New York, 1976, Vol. II. (7) Bulanin, M. O.,J.Mol. Struct. 1986, 141, 315. (8) Almond, M. J.; Downs, A. J.; Perutz, R. N., Inorg. Chem. 1985, 24, 275. (9) Bogdan, P. L.; Wells, J. R.; Weitz, E., J. Am. Chem. Soc. 1991, 113, 1294. (10) Tuan, D. F-T.; Hoffman, R.,J.Am. Chem. Soc. 1985, 24, 871 (11) Perutz, R. N.; Turner, J. J.,J.Am. Chem. Soc. 1975, 97, 4791 (12) Fairhurst, E. C.; Perutz, R. N., Organometallics 1984, 3, 1389. (13) Simpson, M. B.; Poliakoff, M.; Turner, J. J.; Maier, W. B.; McLaughlin, J. G., J. Chem. Soc. Chem. Comm. 1983, 1355. (14) Ishikawa, Y.; Hackett, P. Α.; Rayner, D. M., J. Phys. Chem. 1988, 92, 3863 (15) Rerek, M. E.; Basolo, F.,J.Am Chem. Soc. 1984, 106, 5908. (16) Weiller, Β. H.; Wasserman, E. P.; Bergman, R. G.; Moore, C. B., submitted for publication. The measured activation energy is 2.8 kcal/mol. It should be noted that for the associative mechanism, the activation energy is only a lower bound on the binding energy. (17) Temperature dependent values for the Xe density were taken from: Theeuwes, F.; Bearman, R. J., J. Chem. Thermodynamics 1970, 2, 501. (18) Atkins, P. W., "Physical Chemistry," Oxford University Press, 1978. (19) Ryther, R. J.; Weitz, E., J. Am. Chem. Soc. 1991, 95, 9841 and references contained therein. (20) Rossi, Α., Kochanski, E.; Veillard, Α., Chem. Phys. Lett. 1979, 66, 13. (21) Rettner, C. T., Bethune, D. S.; Scheizer, Ε. K., J. Chem. Phys. 1990, 92, 1442. (22) Dresser, M. J.; Madey, T. E.; Yates, J. T., Surface Sci. 1974, 42, 533-551. RECEIVED

January 28, 1993

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

Chapter 14

Multiphoton Dissociation of Nickelocene for Kinetic Studies of Nickel Atoms Carl E. Brown, M . A. Blitz, S. A. Decker, and S. A. Mitchell

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch014

Steacie Institute for Molecular Sciences, National Research Council of Canada, 100 Sussex Drive, Ottawa, Ontario K1A 0R6, Canada

The mechanism of nickel atom production by multiphoton dissociation of nickelocene at 650 nm is discussed with reference to the population distribution of nickel atoms among metastable excited states, and implications for kinetic studies of chemical reactions of ground state nickel atoms. A mechanism is proposed that involves direct dissociation of nickelocene on an excited state potential surface following absorption of three photons. Rate constants for collisional relaxation of metastable states are shown to correlate with electronic configuration of the excited state, in a manner consistent with qualitative considerations based on model potential curves for the complex NiCO.

Little is known about the gas-phase chemical properties of transition metal atoms. A useful method for producing metal atoms for studies of reaction kinetics near room temperature is laser induced multiphoton dissociation (MPD) of an organometallic precursor. We have investigated a variety of association reactions of transition metal atoms with simple molecules near room temperature, using M P D of volatile metal complexes for production of metal atoms in a static pressure reaction cell ( i ) . Complexes of transition metal atoms with simple molecules are intriguing molecular species that are of interest as simple models for metal-ligand interactions in organometallic and surface chemistry. For kinetic studies of metal atom reactions, details of the M P D process are unimportant, provided that production of metal atoms is effectively instantaneous, and thermalization of metastable excited-states of the atoms occurs rapidly compared with the ground-state chemical reaction of interest. In most cases that we have investigated these requirements are satisfied. However, certain reactions of nickel atoms produced by M P D of nickelocene at 650 nm are exceptional in this regard. This is due to the high reactivity of nickel atoms, coupled with the production of metastable excitedstates which undergo relatively slow collisional relaxation processes. In this chapter we describe studies of MPD of nickelocene at 650 nm, which were undertaken to clarify

0097-6156/93/0530-0177$06.00/0 Published 1993 American Chemical Society

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

178

LASER CHEMISTRY OF ORGANOMETALLICS

the role of metastable state relaxation in the reaction kinetics of nickel atoms with unsaturated molecules such as ethylene (2).

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch014

Production and Monitoring of Nickel Atoms Production of nickel atoms was by pulsed visible laser photolysis of nickelocene at 650 nm, using a Lumonics model EPD-330 dye laser operating on D C M laser dye (Exciton), pumped by a Lumonics model 860-4 XeCl excimer laser. The temporal width of the photolysis pulse was «10 ns. The beam was roughly collimated («1.5 χ 0.5 cm cross section), passed through a NRC model 935-5 variable attenuator and then focussed into the reaction cell using either a 25 cm or 50 cm focal length lens. Relative photolysis pulse energies were measured by directing a fraction of the beam onto a vacuum photodiode. Absolute pulse energies were measured using a Scientech energy meter. Resonance fluorescence from nickel atoms was excited by a probe laser pulse from a second, independently triggered Lumonics excimer laser pumped dye laser, operating on PTP laser dye (Exciton). The photolysis and probe laser beams were collinear and counterpropagating. Laser induced fluorescence (LIF) was separated from scattered photolysis light by a 10 cm focal length monochromator (16 nm F W H M bandpass), and detected by a gated photomultiplier tube. Emission spectra produced by probe and/or photolysis laser excitation were recorded by scanning the monochromator. The pulsed output of the photomultiplier was amplified and then averaged using either a sample and hold circuit and digital voltmeter for computer storage, or by a boxcar integrator for display on a chart recorder. The time dependence of the atom population following the photolysis pulse was observed by scanning the time delay between the photolysis and probe laser pulses, and averaging the fluorescence signal over 10-60 laser shots at each increment of delay. Details of the experimental arrangement have been described elsewhere (1,3). For recording multiphoton ionization (MPI) spectra, a coaxial geometry ionization cell was used (4). Nickelocene vapor above the crystal (Alfa Products) was admitted to the fluorescence or MPI cell using a metal vacuum line. The pressure of nickelocene in the cell was «5 mTorr. Nickelocene was found to be relatively stable as the crystal stored under vacuum, and in the vapor phase, and was convenient to use as a precursor for nickel atoms. Population Distribution of Nickel Atoms. In Figure 1, an energy level diagram of atomic nickel is shown, that includes all states below 15,000 c m excitation energy (5). On the left hand side of the diagram is shown a numbering scheme that is used in the following as a shorthand notation. In this scheme the first number represents a multiplet such as d s F , numbered consecutively from the ground state, and the second number specifies a particular J level within the multiplet. The ground state is thus 11. A l l of the states shown in Figure 1 have electronic transitions that are convenient for fluorescence excitation (6). The transitions that were used to monitor these states are listed in Table I. Strong signals were observed for all states up to and including 31. -1

8

2

3

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

14.

BROWN ET AL.

MPD of Nickelocene for Kinetic Studies of Nickel Atoms 179

T A B L E I: Relative Populations of Metastable States of Ni Following MPD of

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch014

Nickelocene at 650 nm (a) exc

State

X /nm

UF

F

Pop_

11

339.10

1.00

2.0

1.00

21

339.30

0.73

2.6

0.94

22

344.63

0.42

2.0

0.42

12

341.35

0.30

2.6

0.39

23

342.37

0.15

2.3

0.17

13

338.09

0.10

2.6

0.13

31

338.06

0.09

2.9

0.12

e x c

(a) See Figure 1 for notation used for Ni states. X is the wavelength used to excite fluorescence, and LIF is the saturated fluorescence intensity, normalized to that for the ground state transition. Details of the transitions are given in ref. (6). F is a proportionality constant between relative fluorescence intensity and relative population of the lower state before the fluorescence excitation laser pulse. It includes electronic degeneracies of the lower and upper states of the transition, and the quantum yield for fluorescence in the bandpass of the detection monochromator (76). Pop is the relative population of the state following MPD of nickelocene at 650 nm (see the text).

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch014

180

LASER CHEMISTRY O F ORGANOMETALLICS

However, no signals were observed for 41 or 51, which indicates that these states were negligibly populated following photolysis at 650 nm. Estimates of relative populations were made from measurements of fluorescence intensities in the following way. For each transition investigated, the probe laser pulse energy was adjusted so that the transition was saturated with respect to the excitation radiant flux. Under these conditions, the upper and lower states of the transition are populated in the ratio of their electronic degeneracies, so a simple relationship exists between the excited state population at the end of the laser pulse and the total initial population. Knowledge of quantum yields for fluorescence within the bandpass of the detection monochromator thus allows estimates to be made of initial relative populations of the lower states. The required quantum yields are available from tabulated spontaneous transition probabilities (6). Our results are summarized in Table I. Uncertainties in the relative populations are estimated to be ± 30%, based on variations between measurements for the same state using different transitions. Systematic errors may have arisen from involvement of secondary laser excitation processes from excited states of the transitions investigated, or possibly from probe laser photolysis of a molecular fragment of nickelocene produced by the photolysis laser. The photolysis conditions used for these measurements were 0.75 mJ pulse energy at 650 nm, focused with a 25 cm focal length lens. Similar results were obtained when a 50 cm focal length was used. The total pressure in the cell was «5 mTorr, and the time delay between the photolysis and probe laser pulses was «80 ns. These conditions ensured that the population estimates were unaffected by collisions. The relative populations in Table I are shown in the form of a Boltzmann plot in Figure 2. The electronic temperature is seen to be «3000 K. This is consistent with our observation of negligible population in the states 41 and 51. Emission spectra excited by the photolysis laser in the absence of the probe laser showed extremely weak features near 340 nm, most likely due to highly excited nickel atomic photofragments. The weakness of these emissions indicates that such excited nickel atoms were produced in very low abundance. Attempts to observe ionization signals produced by the photolysis laser alone were unsuccessful, although very large signals could be generated by tuning the tightly focused probe laser beam through nickel atomic resonance lines. This shows that multiphoton ionization occurred to a negligible extent at 650 nm, even under tightly focused excitation conditions. MPD of nickelocene was found to occur readily at «445 nm and «555 nm. However, photolysis at these wavelength was not convenient for kinetic studies of nickel atoms because of the occurrence of photochemical depletion of nickelocene in the reaction cell. We attribute this to low photon order photofragmentation processes which deplete nickelocene in a large fraction of the laser beam volume. Photolysis at 650 nm induced MPD only in the small focal region of the photolysis beam. Dependence on Photolysis Fluence. The dependence of the LIF signals for the 21 and 13 states of Ni on the photolysis laser pulse energy at 650 nm is shown in Figure 3. For these measurements the photolysis beam was focused with a 50 cm focal length lens, the total pressure in the cell was «5 mTorr and the delay between the photolysis and probe laser pulses was 80 ns. The 21 signal produced at the lowest photolysis

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch014

14. BROWN ET AL.

MPD of Nickelocene for Kinetic Studies of Nickel Atoms

Figure 1. Low lying electronic states of atomic nickel. Electronic configurations and term symbols are given for the lowest energy member of each multiplet. Bold numbers at left are used as a shorthand notation for the electronic states.

Figure 2. Boltzmann plot showing population distribution of excited states of Ni following MPD of nickelocene at 650 nm. Points are labeled with the state notation of Figure 1.

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

181

182

LASER CHEMISTRY OF ORGANOMETALLICS

energy («30 uJ) is at the limit of detection of our apparatus, and therefore represents a threshold for atom production. At low energy, the signal scales with the pulse energy to the power of 3.6, and shows the onset of saturation at higher energy. The difference between the low energy slopes of the 21 and 13 plots in Figure 3 is not considered to be significant. Mechanism of Nickelocene M P D at 650 nm A schematic diagram that summarizes information on dissociation and ionization energetics of gaseous nickelocene (NiCp2) is shown in Figure 4. The average bond dissociation energy for formation of Ni + 2Cp (Cp = cyclopentadienyl radical) at 298 Κ has been determined from calorimetry (7) as 71.9 kcal mol" , with an uncertainty of ± «3 kcal mol" . Appearance potential measurements (8) give 103 kcal mol" as an upper limit for the energy required to form NiCp + Cp. The ionization energy of N i C p has been determined as 6.29 eV (9). 1

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch014

1

1

2

It is seen in Figure 4 that a minimum of three 650 nm photons is needed to induce dissociation of NiCp . At the three photon level it appears that only the NiCp + Cp channel is possible. This is consistent with the photolysis fluence dependence illustrated in Figure 3, which indicates that the predominant mechanism of atom formation involves more than three photons. Absorption of a fourth photon by N i C p would bring the internal energy well above the ionization threshold, as Figure 4 shows. However, since ionization occurred to a negligible extent under our photolysis conditions, there is an indication that dissociation to NiCp + Cp occurred more rapidly than absorption of a fourth photon by NiCp . The ratio of ionization to neutral dissociation was estimated to be less than 10" . The rate of the dissociation process NiCp -> NiCp + Cp following absorption of three 650 nm photons was estimated using R R K M unimolecular reaction theory (10). The assumption was made that dissociation occurs on the ground state potential surface, following internal conversion of electronic excitation energy to ground state vibrational energy. The vibrational frequencies of NiCp were taken to be those of ferrocene (77), and a loose transition state was assumed in which the vibrational frequencies of the NiCp and Cp fragments were only slightly different from their counterparts in NiCp . The microcanonical rate constant k(E) was found to be less 2

2

2

3

2

2

2

1

1

than 1 s" , and increased to only «100 s" when the first dissociation energy of N i C p 1

2

was assumed to be 85 rather than 103 kcal mol" . The lower dissociation energy is considered to be a lower limit based on thermochemical results for ferrocene (72). These calculations show clearly that, if three photon dissociation to NiCp + Cp occurs as suggested above, then it is a direct rather than a statistical process, involving dissociative excited electronic states rather than internal conversion of electronic to vibrational energy. The estimated statistical dissociation rate is much too slow to account for atom production on the time scale of our experiments, and could not compete with further photon absorption during the laser pulse. The R R K M model was

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

MPD of Nickelocene for Kinetic Studies

BROWN E T A L .

PHOTOL.

Nickel Atoms

ENERGY

10000

•S 1000


ο

NiCp

4-

Cp

30

λ = 650 nm 20

10

NiCp,

Figure 4. Dissociation and ionization energetics of nickelocene.

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

184

LASER CHEMISTRY OF ORGANOMETALLICS

seen to be reasonable in that it predicted thermal dissociation rates near 1250 Κ in rough agreement with experimental values for ferrocene (72). The simplest MPD mechanism consistent with the above is three photon production of NiCp followed by single photon dissociation of NiCp. The near Boltzmann distribution of excited state populations of nickel atoms shown in Figure 2 is consistent with the occurrence of one predominant MPD process of relatively low order. In particular, the observation of negligible populations for the 41 and 51 states of nickel argues against a process involving more than four photons. The threshold photolysis pulse energy for production of nickel atoms at 650 nm is remarkably low. From Figure 3, the threshold for a 50 cm focal length lens is «30 uJ, which corresponds to a peak power of «4 χ 10 Watt cm" . The occurrence of MPD at relatively low intensity is undoubtedly due to the presence of weak d-d absorption bands of nickelocene, for which the extinction coefficient near 650 nm is ε « 60 M ' cm in isooctane solution (73). A very rough estimate of the probability of photon absorption in a 10 ns laser pulse may be found from the extinction coefficient and laser pulse intensity. We find the probability ranges from «0.1 to «1 over the range of pulse energies in Figure 3. These estimates include the factor 1/3 for spatial averaging (9). From these estimates it appears likely that the saturation behavior at higher pulse energies in Figure 3 arises in part from saturation of the initial, single photon absorption step of the overall MPD process.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch014

7

2

-1

Collisional Relaxation of Metastable Excited States of Nickel Atoms The appearance and removal kinetics of all metastable states up to 31 in Figure 1 were investigated, and collisional relaxation rates were measured for Ar, CO, C 0 and C H quenching gases. In all cases, atom production was instantaneous on the time scale of the laser pulse, although collisional relaxation processes led to rather complicated kinetics following the photolysis pulse. The mechanism of relaxation appeared to involve stepwise transitions between adjacent energy levels. Second-order rate coefficients for quenching by Ar, CO, C 0 and C H at 296 Κ are summarized in Table II. The results in Table II allow an assessment to be made of the extent to which collisional relaxation of metastable excited states of nickel atoms influence the kinetics of ground state atom removal by chemical reaction with CO or C H in A r or C 0 buffer gas (2). Simple exponential decay of the ground state population is expected only if metastable state relaxation occurs much faster than ground state reaction. Since the reaction with C H is rather fast (2), significant departures from exponential decay are observed, and can be understood in terms of the relative populations and relaxation rates in Tables I and II. A n example of a nonexponential decay is shown in Figure 5, for the case of reaction of nickel with the precursor nickelocene in the presence of 40 Torr of A r buffer gas. Two curves are drawn through the data points - a simple exponential and a model function that describes the kinetics of species C i n a A - > B - * C - > D consecutive first-order kinetic scheme. In this scheme A and Β represent the metastable states 13 2

2

2

4

4

2

2

2

4

4

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

2

1

14. BROWN ET AL.

MPD of Nickelocene for Kinetic Studies of Nickel Atoms 185

T A B L E II: Rate Constants for Collisional Relaxation of Metastable States a

of Ni Atoms ( ) 11

State

1

(E/cm" )

Config

Ar

CO2

CO

21

dV

7.6

8.3

>8

>8

2.6

5.1

9.0

15

0.03

1.6

0.88

8.7

6.2

4.2

13

18

0.01

2.8

0.13

6.7

1.2

10

13

24

1

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch014

3

k/10" c m s'

3

(205)

2 4 H

D

dV

22

C

(880) 8

12

d s

(1332)

3

2

F

dV

23 (1713)

8

13

d s

2

(2217) 31

dV

(3410) h) (a) See Figure 1 for notation used for Ni states. Excitation energies above the ground state are shown in parentheses. Config gives the electronic configuration, and, for the lowest energy component of each multiplet, the term designation of the multiplet. Rate constants were determined at 296 K.

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

186

LASER CHEMISTRY OF ORGANOMETALLICS

and 12, respectively, C represents ground state atoms and D represents the product of the Ni + N i C p reaction. This scheme is appropriate because the 12 and 13 states are bottlenecks in the overall relaxation process for Ar buffer gas, as Table II shows. The parameters of the model function are the first-order rates for removal of 12, 13 and ground state atoms, and the initial populations of the metastable states relative to the ground state. The parameters for the curve in Figure 5 are consistent with the rate coefficients and relative populations determined in this work, although it was necessary in order to achieve a good fit to use a slightly higher population for (12 + 23) than is indicated in Table I. Thus the kinetics of ground state atom removal can be modeled satisfactorily in this case. When collisional relaxation is induced by gases other than Ar, the bottlenecks in the relaxation process are less pronounced (see Table II), so the simplified kinetic model just described is not applicable. The rate constants for collisional relaxation in Table II are correlated with the electronic configuration of the metastable state. The correlation is particularly evident for Ar and CO quenching gases. Thus, states that have the configuration d s* are more efficiently relaxed than those with the configuration d s . A simple rationalization of this effect may be given in terms of the properties of potential curves of the complex NiCO that correlate with the separated Ni + CO states listed in Table II. States associated with d s* D j or d ^ for which relaxation is relatively efficient, are attractive in character, and correlate with bound molecular states, whereas those

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch014

2

9

8

9

3

2

1

_j < ζ CD M ω Ll M _J

——I

1

0

1

10 TIME

20 DELAY

1

/

30 /xS

•

1—τ*"-

·

40

Figure 5. Kinetic trace showing removal of ground state Ni atoms by reaction with nickelocene («5 mTorr) in presence of 40 Torr A r buffer gas at 296 K. Effects due to collisional relaxation of metastable excited state Ni atoms are shown by fitted curves (see the text).

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

14.

BROWN ET AL.

8

MPD of Nickelocene for Kinetic Studies of Nickel Atoms

187

2 3

associated with d s F j are thought to correlate with states of NiCO that are repulsive at intermediate Ni-CO separation (14). It is likely that these properties of the d s* and d s configurations apply for other Ni atom - molecule interactions, since they arise from the ability of the d s* configuration to minimize σ-repulsion by polarization of selectron density away from the molecule (75). Reduced repulsive interactions allow closer collisions with enhanced probabilities for energy transfer by curve crossing processes. It is noteworthy that electron spin appears to be a much less important factor than electronic configuration. This is shown by the observation that the relaxation efficiency of the 31 state ( d s* ^D?) is in all cases relatively large, even though this state can only relax by a transition to a state of different spin multiplicity. This is consistent with the observation that the formally spin-forbidden association reactions of nickel atoms with CO and C H occur readily at room temperature (2). 9

8

2

9

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch014

9

2

4

Literature Cited 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Mitchell, S. A. In Gas-Phase Metal Reactions; Fontijn, Α., Ed.; Elsevier: Amsterdam, 1992, Chapter 12; pp 227-252. Brown, C. E.; Mitchell, S. Α.; Hackett, P. A. Chem. Phys. Lett. 1992, 191, 175. Parnis, J. M.; Mitchell, S. Α.; Hackett, P. A. Phys. Chem. 1990, 94, 8152. Mitchell S. Α.; Hackett, P. A. J. Chem. Phys. 1983, 79, 4815. Corliss, C.; Sugar, J. J. Phys. Chem. Ref. Data 1981, 10, 197. Fuhr, J. R.; Martin, G. Α.; Wiese, W. L.; Younger, S. M. J. Phys. Chem.Ref. Data 1981, 10, 305. Pilcher, G.; Skinner, H. A. In The Chemistry of the Metal-Carbon Bond; Hartley, F. R.; Patai, S., Eds.; Wiley: New York, 1982, Vol. 1; pp 43-90. Hedaya, E. Accounts Chem. Res. 1969, 2, 367. Leutwyler, S.; Even, U.; Jortner, J. Chem. Phys. 1981, 58, 409. Gilbert, R. G.; Smith, S. C. Theory of Unimolecular and Recombination Reactions; Blackwell Scientific Publications, Oxford: 1990. Aleksanyan, V. T. In Vibrational Spectra and Structure; Durig, J. R., Ed.; Elsevier: Amsterdam, 1982, Vol. 11, pp 107-167. Lewis, Κ. E.; Smith, G. P. J. Am. Chem. Soc. 1984, 106, 4650. Gordon, K. R.; Warren, K. D. Inorg. Chem. 1978, 17, 987. Bauschlicher, C. W. Jr.; Barnes, L. Α.; Langhoff, S. R. Chem. Phys. Lett. 1988, 151, 391. Bauschlicher, C. W. Jr.; Bagus, P. S.; Nelin, C. J.; Roos, B. O. J. Chem. Phys. 1986, 85, 354. Mitchell, S. Α.; Hackett, P. A. J. Chem. Phys. 1990, 93, 7813.

RECEIVED March 31, 1993

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

Chapter 15

Ionization and Fluorescence Techniques for Detection and Characterization of OpenShell Organometallic Species in the Gas Phase 1

2

1

William R. Peifer , Robert L. DeLeon , and James F. Garvey

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch015

1

Department of Chemistry, State University of New York at Buffalo, Buffalo, NY 14214 Physical and Chemical Sciences Department, Arvin/Calspan Corporation, P.O. Box 400, Buffalo, NY 14225

2

Recent advances are discussed in the development of electronic spectroscopic probes for the study of excited-state structure and photodissociation dynamics of gas-phase organometallics. Because of the short timescale for intermolecular energy transfer within van der Waals clusters, the UV photodissociation dynamics of cluster-bound transition metal carbonyls differs considerably from the photodissociation dynamics of the naked species in the gas phase. It is therefore possible to employ multiphoton ionization to produce cluster-bound metal carbonyl photoions in high yield. Resonant photoionization (accomplished with tunable lasers) and mass-resolved detection allow one to probe the excited states of both closed-shell and open-shell neutral organometallics. Finally, a time-resolved two-laser technique employing fluorescent detection of atomic photoproducts is described. This technique allows one to study photodissociation dynamics of organometallics with a temporal resolution competitive with the fastest transient absorption techniques, and a level of sensitivity which is far superior. Coordinatively unsaturated organotransition metal species play a central role in mechanistic organometallic chemistry. Such species are important both in stoichiometric reactions (7,2) and, perhaps more importantly, in a wide range of catalytic reactions: for example, synthetically important hydrogénation reactions (3); industrially important hydroformylation (4) and polymerization reactions (5); automotive engine combustion control and emission reduction (6); and biologically important processes such as nitrogen fixation (7) and oxygen transport (8). It has been suggested that the sin0097-6156/93/0530-0188$06.00/0 © 1993 American Chemical Society

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch015

15.

PEIFER ET AL.

Open-Shell Organometallic Species in the Gas Phase

gle most important property of a homogeneous catalyst is a vacant coordination site(9). Clearly, the production of model coordinatively unsaturated species in the laboratory, and the development of fast, sensitive probes for detection and characterization of these species, impacts upon a broad scope of disciplines in the natural sciences. A convenient route to the synthesis of coordinatively unsaturated molecules in the laboratory is the photolysis of appropriate saturated precursors by pulsed visible or U V laser irradiation. While single-quantum photodissociation in condensed phases is characterized exclusively by the breaking of a single bond and the production of a single site of coordinative unsaturation (10), photodissociation of a gas-phase molecule may lead to the breaking of several bonds (depending on the energy of the photon absorbed) and the production of multiple vacancies in the coordination sphere (77). These coordinatively unsaturated molecules are highly reactive: subject to the constraints of electronic spin conservation, they will undergo ligand recombination reactions in the gas phase at or near the gas-kinetic limiting rate (72), and will undergo coordination with solvent (or host) molecules in liquids and frozen matrices at diffusionlimited rates (13-19). In order to detect and characterize such highly reactive species, one must employ techniques which are sufficiently sensitive to allow detection at low number densities, and sufficiently fast to avoid interference from competing solvation and/or ligand recombination reactions. Available Spectroscopic Probes for Organometallics Vibrational Probes. One powerful method by which coordinatively unsaturated species have been identified and studied is time-resolved IR absorption spectroscopy, or TRIS (20). Such experiments are typically performed by pulsed U V laser photodissociation of an appropriate saturated precursor in a gas cell. Transient changes in the attenuation of a continuous infrared beam passed down the length of the cell, due to disappearance of the saturated precursor and appearance of the various photoproducts following U V photolysis, are monitored by a fast IR detector. Primary photoproduct identities are usually assumed on the basis of results from gas-phase chemical trapping studies (77) as well as known transition frequencies from matrix isolation studies (27), while photoproduct geometries can often be inferred from group theoretical analysis of the number and relative intensities of IR absorption features (22). While the application of TRIS has led to advances in our understanding of the structure and reactivity of coordinatively unsaturated organometallic species (primarily those in the electronic ground state), the technique is limited by a few significant constraints. First, because this is an absorption technique, the product of the number density and extinction coefficient of a target molecule must be large in order to allow for the molecule's detection. The large extinction coefficients associated with C-O stretching modes of metal carbonyls allow for detection of these species at low num-

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

189

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch015

190

LASER CHEMISTRY OF ORGANOMETALLICS

ber densities, but weaker absorbers must be present at higher number densities. For sufficiently weak absorbers, required number densities are large enough that reactive bimolecular collisions may take place during the course of product detection, thus complicating spectral interpretation. Second, because one employs infrared detectors, the temporal resolution of transient experiments is limited. Typical detector/preamp combinations for this type of experiment display simple exponential risetime constants in the range from 100 to a few hundred nanoseconds (23). The microscopic details of processes which take place on a much shorter timescale cannot be directly probed. Finally, since TRIS is a purely vibrational spectroscopic probe, one is unable to distinguish directly between vibrational transitions taking place in the ground electronic manifold and those taking place in an excited electronic manifold of the target molecule. The ability to make such a distinction is desirable, since one would expect that: (a) the excited states of coordinatively unsaturated organometallic species would represent a significant fraction of total photoproducts in some photophysical schemes, and (b) the structure and reactivity of electronically excited states of these species would differ significantiy from those of the ground state. Clearly, a spectroscopic technique which probes the electronic structure of coordinatively unsaturated organometallic species would allow the experimentalist to acquire information complementary to that provided by vibrational probes such as TRIS. However, interrogation of the electronic structure of these species is thwarted by the very property which makes them easy to produce and interesting to study in the first place; that is, the photolability of coordination compounds, both saturated and unsaturated. Electronic surface crossings among excited states of these molecules occur at rates which are often faster than those of optical transitions. Electronic relaxation generally occurs via non-radiative mechanisms and is often accompanied by the breaking of one or more metal-ligand bonds. Consequendy, with a few notable exceptions (24-26), powerful techniques such as laser induced fluorescence (LIF) and multiphoton ionization (MPI) have not enjoyed general applicability to the study of electronic excited states of coordinatively unsaturated organometallic species. Electronic

Probes.

The photodissociation behavior of organometallic molecules is clearly governed by the interplay between intramolecular energy partitioning, and intermolecular energy transfer. The photodissociation dynamics of these species evolves from behavior characteristic of the naked molecule (where energy disposal via bond cleavage predominates), to that characteristic of solvated molecules (where energy disposal via solvent heating predominates). By studying this behavioral evolution experimentally, one can in principle derive detailed insights about the electronic structure of these transient excited species. Weakly bound van der Waals clusters represent a convenient, finite experimental model of the complex, virtually infinite environment of condensed matter. Such clusters can be easily formed by expanding a buffer gas such as helium,

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch015

15.

PEIFER ET AL.

Open-Shell Organometallic Species in the Gas Phase

seeded with a small proportion (perhaps a fraction of a percent) of a compound of interest, into a vacuum chamber through an orifice of small diameter. The cluster beam thus produced is amenable to interrogation by a broad array of spectroscopic techniques. An area of ongoing research in our laboratory at SUNY, Buffalo, is the photodissociation dynamics of organometallics bound within van der Waals clusters. The ultimate goal of this research is to provide a better understanding of excited electronic states of organometallic molecules, particularly those which are coordinatively unsaturated, and the roles of these excited states in mediating energy transfer. It is hoped that a better understanding of these phenomena will lead to the development of novel approaches for fast, sensitive detection and characterization of this important class of molecules. Photodissociation Dynamics of Homogeneous Metal Carbonyl van der Waals Clusters Fe(CO)5 van der Waals Clusters. Perhaps the earliest indications of the unusual photodissociation dynamics of cluster-bound organometallics came out of an investigation of the photoionization behavior of jet-cooled Fe(CO)5 by Smalley and coworkers (27). Jet-cooled monomer from a seeded helium expansion was observed to undergo MPI following irradiation at 157 nm by an F2 excimer laser, giving rise to the expected photoions (i.e., a very strong signal due to atomic metal ions, and very weak signals due to the iron mono- and dicarbonyl ions). At this wavelength, neutral Fe(CO)5 clusters larger than the trimer were observed to undergo one-photon ionization, giving rise to a distribution of cluster ions of the type, [Fe(CO)5] . Drastically different results were obtained following irradiation at 193 nm from an ArF excimer laser. At this wavelength, two sequences of cluster ion signals were observed by time-of-flight mass spectrometry: a major sequence, attributed by Smalley and co-workers to bare metal cluster ions, F e ; and a minor sequence, attributed to F e ( C O ) . Because ^°Fe (natural abundance 91.8%) has an atomic mass indistinguishable from twice the molecular mass of ^ C ^ O (natural abundance 98.7%), it was not possible for Smalley and co-workers to assign a unique stoichiometry to any given cluster ion. They reasoned that since cluster ions containing odd numbers of CO ligands (and therefore falling between successive F e clusters) constitute only a small fraction of the total photoion yield, cluster ions containing even numbers of CO ligands probably constitute only a minor degree of isobaric interference. Photoion yields were found to be highly wavelength dependent; for example, MPI of the cluster beam at 266 nm failed to produce any significant cluster photoion signal. The photodissociation and ionization behavior of these van der Waals clusters is rather remarkable, since the photoproduct ion distribution implies extensive rearrangement of both strong (covalent) and weak (dispersive) bonds within the clusters following pho+

n

+

x

+

x

v

+

x

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

191

192

LASER CHEMISTRY O F ORGANOMETALLICS

toexcitation, and the pronounced wavelength dependence is observed over a narrow spectral region where absorption is diffuse. M o ( C O ) ^ van der Waals Clusters. Investigations of metal carbonyl van der Waals cluster photophysics in the SUNY laboratory began in late 1988 (28). Mo(CO)^ was chosen for initial experiments for two reasons. First, much was already known about the one-photon photodissociation dynamics of the Group VIB hexacarbonyls in both gas and condensed phases, offering a body of existing knowledge upon which to build. Second, although the photodissociation dynamics of clusters of Fe(CO)5 * certain amount of attention, clusters of homoleptic metal carbonyls from other triads had not yet been examined. Given the ready availability of the hexacarbonyls of chromium, molybdenum, and tungsten, it might then be possible to study the systematics of intracluster photochemistry in a series of analogous metal carbonyls. Although the quadrupole mass analyzer originally employed for photoproduct ion detection in the pulsed molecular beam apparatus was limited in terms of sensitivity and mass range, it enabled the collection of both electron-impact and MPI mass spectra of the Mo(CO>6 cluster beam.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch015

n a (

r e c e r v e o 1

a

Mass Spectroscopic Analysis of the M o ( C O ) ^ Cluster Beam. Mass spectra of the Mo(CO)^ cluster beam collected following either electron impact or multiphoton ionization are shown in Figure 1. The electron impact mass spectra for the cluster beam and for an effusively introduced sample of Mo(CO)^ were indistinguishable over the m/z range illustrated. The signals observed correspond to the daughter ions expected following prompt statistical decay of nascent Mo(CO)6 parent ions. The MPI mass spectrum of naked Mo(CO)^ is dominated by atomic metal ions and, as expected, displays no signals due to molecular ionization. The MPI mass spectrum of the Mo(CO)£ cluster beam, however, displays totally unexpected features, which were attributed to the formation of metal oxide ions. +

Possible Origins of the Metal Oxide Photoions. What is the mechanism by which these ions arise, and what inferences may be derived regarding cluster photophysics? From the dependence of the metal oxide ion signal upon the relative timing of the photoionizing laser and the molecular beam pulse, and from the absence of metal oxide ions in the MPI spectrum of naked Mo(CO)^, we deduced that these unexpected photoions were arising from species present in the molecular beam rather than from oxygen-containing impurities in the vacuum chamber. However, no oxygen-containing species other than the carbonyl ligands themselves could be detected in the molecular beam gas mixture. This implied that some unusual bimolecular chemical reaction was taking place among metal carbonyl species in the molecular beam. Reactions occurring fast enough to reach completion within the ca. 20-nanosecond duration of the laser pulse would be most likely if they were occurring between

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

15.

Open-Shell Organometallic Species in the Gas Phase

PEIFER ET AL.

A

MofcO)

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch015

• — ρ

MoCO*

Mo(CO)

+

193

+ 3

1

2

ι—> x l O Mo

Β

i.

+

:, ;—ι

i" * '

A*

1

• ':.

•

Î i

Jf * U j c

>

x

10

Mo* '

Μ

•

ι

M

| '

·: r- ; ·

ι

MoO*

0 0

*! !

ι Η.

Π*&ι

« .;Ι|.' ι

:;·· 111 kcal/mol, which is remarkable. The lanthanide contraction and spin-orbit (relativistic) effects lead to changes in orbital size and stability which increase bonding overlap and decrease loss of exchange stabilization on bond formation. The unusual ability of the 5dseries M+ to dehydrogenate C H is apparently a direct result of their unique electronic structure. +

0

2

4

2

M + + Hydrocarbon Reactions The second ionization potentials of transition metal atoms lie in the range 12-20 eV, which is larger than the first ionization potential of all alkanes and alkenes except C H . Dipositive metal cations should be extraordinarily strong oxidizing agents capable of removing an electron even from alkanes. When we began studying reactions of T i with alkanes, we expected to observe facile electron transfer followed by explosion of alkane into fragments. In fact, M ions are indeed highly reactive, but the chemistry is often remarkably gentle. In a fast flow reactor at 0.4 torr He, we observed the following (19): 4

2 +

+

2

+

2

Ti + + η C H

2

Ti(CH ) +, η = 1-4

4

4

n

2

Ti + + C H - * T i H + + C H + 2

6

2

5

(adduct stabilization)

(4)

(hydride anion transfer)

(5)

(electron transfer)

(6)

2

Ti + + C H ^ T i + + C H + . 3

8

3

8

2

Ti + is formed by laser vaporization of a rotating T i disk upstream of the flow tube; rate constants are measured by mass spectrometric detection of the decay of Ti + vs calibrated flow of alkane reactant at fixed reaction length. A l l three reactions are fast; efficiencies k/k are ~0.1 for T i + C H and - 1 for T i + C H and C H . Here k is the Langevin rate constant for cation-neutral capture collisions. The second ionization potential of T i is 13.6 eV, so electron transfer is exothermic from all three alkanes. Nevertheless, the collisionally stabilized dipositive adduct ions Ti(CH ) + dominate Rxn. 4, surviving - 1 ms in a 300 Κ 2

2 +

L

2

6

3

8

2 +

4

L

2

4

n

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

16.

213

Reactions of Gas-Phase Transition-Metal Atoms

WEISSHAAR

He bath. We observe no electron transfer in either the C H or C H reactions. In Ti + + C H , hydride transfer dominates. Electron transfer is observed only in Ti + + C H . A simple one-dimensional curve-crossing model (Figure 3) can explain the remarkable specificity of these reactions. At long range, reactants T i + RH follow the attractive ion-induced dipole potential V ^ r ) = - aq /2r , where a is the R H polarizability and q = +2 is the ion charge. The T i R H adduct-ion potential well may lie below the ion-pair asymptotes. Bimolecular ion-pair products Ti+ + RH+ and TiH+ + R+ follow repulsive Coulomb potentials V (r) = +q /r at long range, crossing the T i + R H curve at a radius that depends sensitively on the ion-pair product's exothermicity. The different alkanes C H , C H , and C H in effect tune the radii of the crossing points between reactants and the two ion-pair product channels, thus controlling the product branching. Known reaction exothermicities and the simple Coulomb potentials allow us to calculate the radii of the crossing points. The calculated electron-transfer curve crossing occurs at 14.4 Â for T i + C H , 7.0 À for C H , , and 6.1 À for C H . Apparently the electron will not jump at 14.4 À or 7.0 A , but it jumps efficiently at 6.1 À, so electron transfer dominates the C H reaction products. The calculated hydride transfer curve crossing occurs at 9.1 À for C H , 4.6 À for C H , and 4.5 A for C H . T i + C H reactants survive the electron transfer crossing and reach r = 4.6 À, where Htransfer occurs and products separate. Ti + + C H survives both the electron and hydride transfer crossings, reaching the adduct ion well at short range. About 10% of the adducts are collisionally stabilized at the He pressure of 0.4 torr. The adducts may be thermodynamically stable relative to bimolecular products, or they may be metastable but separated from bimolecular products by a large barrier. The remaining adducts dissociate back to Ti + + C H , oblivious to the electron transfer and hydride transfer crossings on the way out as they were on the way in. Thus the specificity of the Ti + data strongly suggests that both eand H- transfer are "physical" processes controlled by the radii of long-range curve-crossing points, which are in turn determined by the exothermicities of these two channels. Given the simplicity of the Ti + chemistry, we were quite surprised when Freiser and co-workers (20) found an array of M + alkane reactions that produce the same kinds of H and C H elimination products observed in M+ + alkane reactions (see Rxns. 1-3). The combined data suggest the following simple model. As M approaches the alkane, it always encounters the e- transfer curve crossing first; the efficiency of e- transfer depends critically on the radius of the crossing. Those M + alkane pairs that survive the e- transfer crossing may encounter a H- transfer crossing at 4-6 À; if so, H- transfer is efficient. But if the calculated H- transfer radius is smaller than about 4 À, H- transfer does not occur. Other rearrangement channels leading to H or C H elimination then dominate the products. Apparently the chemical forces become too strong at short range to allow M + to escape with H- attached. Remarkably, gas phase M + and M+ may be rather similar chemicals when the M survives both the electron and the hydride transfer crossings at long range and gains close contact with the alkane. If anything, M at short range appears even better able to break C-H and C-C bonds than M+. This leads us to wonder if neutral transition metal atoms would be active chemicals as well, if only they were able to penetrate long-range potential barriers to bond insertion. 4

2

6

2

2

6

2

3

8

2 +

2

4

2 +

2

2

2 +

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch016

4

2

6

3

8

2 +

4

2

3

3

8

8

2 +

4

2

6

3

8

2

2

4

2

4

2

2

2

2

+

4

2 +

2

2

4

2

2

2

+

2

+

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

+

6

LASER CHEMISTRY OF ORGANOMETALLICS

214

M + Hydrocarbon Reactions Overview of Kinetics Measurements. With notable exceptions, study of the chemistry of neutral gas phase metal atoms has been dominated by oxidation reactions of alkali and alkaline earth metals. Little is known about the chemistry of neutral transition metal atoms. Mitchell and co-workers (21) are systematically studying the kinetics of termolecular association of 3d-series transition metal atoms with small ligands including 0 , N O , C O , N H , C H , and 2

3

2

4

C2H2.

Our own contribution (22) has been a broad survey of the reactivity of transition metal atoms with oxidants, alkanes, and alkenes at 300 K . We use the same fast flow reactor as in the T i kinetics measurements, substituting a hollow cathode sputtering source of metal atoms for the laser vaporization source. Laser-induced fluorescence (LIF) probes the decay of specific electronic states of M vs calibrated reactant flow at fixed reaction length. The slope of plots of ln(M) vs reactant number density yields an effective bimolecular rate constant. Thus far, all of the results are in He buffer gas at 0.4-0.8 torr and 300 K . The experiment is described in more detail in the Ph.D. thesis of David Ritter (23). Table I collects the rate constants for reactions of neutral M from the 3d-series with a variety of collision partners. We find no reaction (k < 1 χ 10" cm -s ) of 3d-series metal atoms with the linear hydrocarbons propane and /z-butane. We observe fairly efficient reactions of Sc, T i , V , and N i with alkenes. Rate constants generally increase with the size of the alkene. They are independent of which spin-orbit level J within the ground term is probed by LIF. We observe no dependence ( ± 2 0 % precision) of the effective bimolecular rate constant on He pressure over the range 0.4-1.0 torr. Either the reactions are bimolecular or they have already reached the saturated regime of termolecular kinetics at 0.5 torr. Our quantitative kinetics measurements for ground states of M at 300 Κ provide upper bounds on activation energies that can test ab initio estimates of barrier heights. For example, the Ni(d s, D) + alkene data imply activation energies not larger than 0.4-3.5 kcal/mol, depending on the alkene. Mitchell and co-workers obtain a N i - C H bond energy of 35.5 ± 5.0 kcal/mol (24). Calculations on N i 4- C H find the crossing point between the repulsive A surface and the attractive A surface (from Ό Ni) to lie - 8 kcal/mol above Ni( D) + C H , perhaps a slight overestimate (25). Most recently, we have begun to extend this work to the 4d-series neutral transition metal atoms (26). Zr is a chemically lively metal which reacts with alkenes at roughly one-half the hard-spheres collision rate at 300 K . Zr also abstracts oxygen atom from 0 with rate constant 5 χ 1 0 cm -molec -s at 300 K . That is some 20 times more efficient than the congener Ti. We have recently found that Mo reacts with both alkenes, in contrast with its congener Cr, which is inert.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch016

2 +

14

3

1

9

2

3

4

3

2

4

1

l

ι

l

3

2

4

11

3

1

1

2

Electronic Structure Considerations. It is already clear that the neutral transition metal atoms are much more inert than the cations M . Neutral M does not react with alkanes, and only Sc, T i , V , and N i react with alkenes. The relative inertness of neutral M is easily explained. Most of the ground states are 3d 4s , and the 4s orbital is much larger than 3d. The traditional picture of bonding between metal and alkene is the Dewar-Chatt-Duncanson model, in which two donor-acceptor interactions occur. Doubly-occupied 3d on the metal +

x2

2

xz

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

16. WEISSHAAR

215

Reactions of Gas-Phase Transition-Metal Atoms

(donor) overlaps the empty n* on ethylene (acceptor) and doubly-occupied π on ethylene (donor) overlaps the empty 4s on the metal (acceptor).

12

1

T A B L E I. Effective bimolecular rate constants ( 1 0 cn^-molec^-sec ) for reactions of 3d-series, ground state neutral metal atoms at 0.80 torr He and 300 K. Rate constants precise to ± 1 0 % and accurate to ± 3 0 % .

Sc D

Ti

NR 9.5 8.3 NR 14 20 33 71 29 NR 0.01 NR

NR 6.2 5.0 NR 7.1 7.5 14 52 10 NR NR NR

Reactant

V 4

Cr S

Mn S

Fe *D

Co 4

Ni F

Cu S

NR 9.6 9.6 NR 14 22 30 68 15 NR 0.02 NR

NR NR

NR NR

NR NR

NR NR

0.5 11 21 0.3 140 160 155 67 110 NR 10 NR

NR NR

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch016

2

ethene propene-h propene-d C F 1-butene t-2-butene c-2-butene isobutene 1,3-butadiene n-butane cyclopropane propane 6

6

3

6

F

7

6

F

-

-

-

-

NR NR NR NR NR 0.15 NR NR NR

NR NR NR NR NR NR NR NR NR

NR NR NR NR NR NR NR NR NR

NR 0.09 NR NR NR 0.35 NR NR NR

14

3

2

3

NR NR NR NR NR NR NR NR NR

1

NR means no reaction observed; k less than 10" cm -s . Dash (--) means reaction not studied.

None of the neutral M atom ground states is properly prepared for a favorable interaction with C H . As M and alkene approach, the metal atom must hybridize to relieve repulsion between 4s or 4s and the doubly occupied π orbital on alkene. Two mechanisms are possible, sd hybridization and sp hybridization; in reality, the optimal combination of these will occur. In sd hybridization, 3d 2 and 4s mix to form two hybrids, which we label sd+ and sd.. One hybrid (sd+) concentrates probability along the z-axis; this becomes the acceptor orbital. The other (sd.) concentrates probability in the xy-plane, thus relieving repulsion; this becomes the singly or doubly occupied non-bonding orbital. In sp hybridization, 4p and 4s mix to form two hybrids directed towards (sp ) or away from (sp.) the approaching C H rr-orbital. The sp hybrid becomes the acceptor and the sp. becomes the lone pair, polarized to the rear of the metal atom to relieve repulsion (27,28). Why are Sc, T i , V , and N i the only reactive 3d-series metal atoms? Compared with Fe and Co, N i has a very small promotion energy to a state of low-spin, 3d 4s character (10 kcal-mol for N i vs 34 for Fe and 21 for Co). 2

4

2

1

z

z

+

2

xl

4

1

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

+

LASER CHEMISTRY OF ORGANOMETALLICS

216

Only such a low-spin state can hybridize and accomodate two non-bonding electrons in the resulting sd hybrid. The avoided intersections of the repulsive surfaces from 3d 4s ground state reactants and the attractive surfaces from lowspin 3d 4s excited state reactants produce barriers to M-alkene adduct formation (Figure 4). The barrier height scales roughly as the energy difference between 3d 4s and low-spin 3d 4s atomic states. These ideas are familiar from the V+ and Fe+ + alkane examples. N i overcomes the barrier at 300 Κ while Fe and Co cannot. The sd-hybridization scheme is preferred in N i because the 4p orbitals lie very high in energy. It is less obvious why Sc, T i , and V react with alkenes at such similar rates. Electron spin permits both the sd- and sp-hybridization scheme to operate from the ground states of Sc, T i , and V . Since the promotion energy to low-spin 3d *4s states varies widely from Sc to Ti to V but the rate constants are quite similar, we suggest participation of 4p orbitals in the bonding. The early data indicate that Zr (4d 5s ground state) is much more reactive than its 3d-series congener T i . The energies of corresponding types of terms are very similar in Ti and Zr. Consequently, we suggest that orbital sizes play an important role. In the 4d series, 4d and 5s orbitals have comparable size; in the 3d series, 3d is much smaller than 4s. Repulsive 4s-alkene interactions and bonding 4d-alkene interactions begin at more similar distances, attenuating the steepness of the repulsive surfaces from 4d 5s . In addition, the larger absolute size of 4d compared with 3d ultimately makes stronger chemical bonds due to better spatial overlap. Mo (high-spin, 4d 5s ground state) reacts with alkenes at 300 K , albeit slowly. In contrast, the 3d-series congener Cr (3d 4s*) is inert. The same orbital size considerations come into play comparing Mo to Cr. In addition, electron spin is a poorer quantum number in the 4d series than in the 3d series. This means that crossings between nominal high-spin and low-spin surfaces will be more strongly avoided, making nominal spin-changing events more facile. n2

2

nl

n2

2

nl

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch016

n_

2

2

x2

5

2

L

5

Summary and Prognosis At a certain level, we are beginning to understand how the electronic structure of transition metal atoms determines their gas phase chemical reactivity. A recurring theme is the importance of intersections among different kinds of diabatic surfaces arising from the closely-spaced atomic asymptotes. It is really the overall pattern of low-lying states that determines reactivity, because the ground state itself is often ill-prepared to form chemical bonds. Perhaps the presently perceived differences among M , M+, and M + chemistry will someday seem superficial. The current data may show that chemical reactivity is governed by the ability of the metal atom to gain close access to the hydrocarbon, which is determined by the long-range part of the metal-hydrocarbon potential. That is, most of these systems may have lowenergy pathways from M(hydrocarbon) at short range to elimination products, but only some of the systems, primarily those with +1 or +2 charge, see small enough barriers to access those pathways. We have only begun to probe the truly subtle chemical questions involving the geometries and stabilities of the intermediates between F e + C H and FeC H + + C H , for example. Spectroscopic work could prove important here. Finally, we can ask the question whether gas phase work is relevant to solution phase organometallic chemistry. There is an enormous literature of C-H bond activation in solution phase. Typically thermolysis or photolysis initiates 2

+

3

8

2

4

4

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch016

16.

WEISSHAAR

Reactions of Gas-Phase Transition-Metal Atoms

F I G U R E 3. One-dimensional model of M2+ + alkane reactions, showing the radii of the electron transfer (r ) and hydride transfer (r*) curve crossings. 1

Ni ( 3 d 9 4 s V D ) + C H 2

4

XT Ni ( 3 d 8 4 s 2 , 3 F ) Ν ΐ Ο , Η ^ )

+

F I G U R E 4. Schematic diabatic potential energy curves describing the approach of neutral Fe, Co, or N i to an alkene. A barrier occurs on the lowest-energy adiabatic surface due to the avoided crossing of the curves.

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

217

218

LASER CHEMISTRY O F ORGANOMETALLICS

the reaction by creating a coordinatively unsaturated metal center. Examples include complexes of Ru and Rh from the 4d series and of Ir and Pt from the 5d series. This is reminiscent of the list of M+ cations that can dehydrogenate methane at low collision energy (Rxn. 3). The same metals would make a good list of heterogeneous catalysts as well. To some extent the highly active metals remember who they are, whether surrounded by vacuum, ligands and solvent, or other metal atoms!

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch016

Acknowledgments. I thank the National Science Foundation and the Donors of the Petroleum Research Foundation for continuing support of our research on the structure and reactivity of transition metal species. M y former and current graduate students Dr. Lary Sanders, Dr. Russell Tonkyn, Dr. David Ritter, Dr. Scott Hanton, Dr. Andrew Sappey, M r . Robert J. Noll, and M r . John Carroll deserve most of the credit for the progress we have made.

Literature Cited 1. P.B. Armentrout in Gas Phase Inorganic Chemistry, edited by D.H. Russell (Plenum, New York, 1989), pp. 1-42; S.W. Buckner and B.S. Freiser, ibid., pp. 279-322; D.P. Ridge and W.K. Meckstroth, ibid., pp. 93-113; R.R. Squires and K.R. Lane, ibid., pp. 43-91; D.H. Russell, D.A. Fredeen, and R.E. Teckleburg, ibid., pp. 115-135; M.F. Jarrold, ibid., pp. 137-192; D.K. MacMillan and M.L. Gross, ibid., pp. 369-401. 2. L.M. Roth and B.S. Freiser, Mass Spec. Rev., in press. 3. For example, see M. Hackett and G.M. Whitesides, J. Am. Chem. Soc. 110, 1449 (1988) and references therein. 4. C.E. Moore, NBS Circ. No. 467 (U.S. Dept. of Commerce, Washington, D.C., 1949, 1952); C. Corliss and J. Sugar, J. Phys. Chem. Ref. Data 14 (Supp. 2) 407 (1985). 5. P.B. Armentrout and J.L. Beauchamp, Acc. Chem. Res. 22, 315 (1989); J. Allison in Progress in Inorganic Chemistry, edited by S.J. Lippard (WileyInterscience, New York, 1986); P.B. Armentrout, Ann. Rev. Phys. Chem. 41, 313 (1990); P.B. Armentrout, Science 251, 175 (1991). 6. G. Ohanessian, M.J. Brusich, and W.A. Goddard III, J. Am. Soc. 112, 7179 (1990). 7. L.F. Halle, P.B. Armentrout, and J.L. Beauchamp, Organometallics 1, 963 (1982); E.R. Fisher, R.H. Schultz, and P.B. Armentrout, J. Phys. Chem. 93, 7382 (1989); R. Georgiadis, E.R. Fisher, and P.B. Armentrout, J. Am. Chem. Soc. 111, 4251 (1989); S.K. Loh, E.R. Fisher, L. Lian, R.H. Schultz, and P.B. Armentrout, J. Phys. Chem. 93, 3159 (1989); R. Georgiadis and P.B. Armentrout, J. Phys. Chem. 92, 7067 (1988); R. Georgiadis and P.B. Armentrout, J. Phys. Chem. 92, 7060 (1988); L.S. Sunderlin and P.B. Armentrout, J. Phys. Chem. 92, 1209 (1988). 8. W.D. Reents, F. Strobel, R.G. Freas, J. Wronka, and D.P. Ridge, J. Phys. Chem. 89, 5666 (1985); F. Strobel and D.P. Ridge, J. Phys. Chem. 93, 3635 (1989). 9. R. Tonkyn and J.C. Weisshaar J. Phys. Chem. 90, 2305 (1986); R. Tonkyn, M. Ronan, and J.C. Weisshaar, J. Phys. Chem. 92, 92 (1988). 10. D.B. Jacobson and B.S. Freiser, J. Am. Chem. Soc. 105, 5197 (1983); G.D. Byrd, R.C. Burnier, and B.S. Freiser, J. Am. Chem. Soc. 104, 3565 (1982).

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch016

16.

WEISSHAAR

Reactions of Gas-Phase Transition-Metal Atoms

219

11. J.C. Weisshaar in Advances in Chemical Physics, edited by C.Ng (WileyInterscience, New York, 1991). 12. P.R. Kemper and M.T. Bowers, J. Am. Chem. Soc. 112, 3231 (1990). 13. P.A.M. van Koppen, J. Brodbelt-Lustig, M.T. Bowers, D.V. Dearden, J.L. Beauchamp, E.R. Fisher, and P.B. Armentrout, J. Am. Chem. Soc. 113, 2359 (1991). 14. L. Sanders, A.D. Sappey, and J.C. Weisshaar, J. Chem. Phys. 85, 6952 (1986); L. Sanders, S.D. Hanton, and J.C. Weisshaar, J. Chem. Phys. 92, 3485 (1990). 15. N. Aristov and P.B. Armentrout, J. Am. Chem. Soc. 108, 1806 (1986); N. Aristov, Ph.D. Thesis, Univ. of California-Berkeley, Dept. of Chemistry (1988). 16. S.D. Hanton, R.J. Noll, and J.C. Weisshaar, J. Chem. Phys. 96, 5165 (1992); J. Chem. Phys. 96, 5176 (1992). 17. R.H. Schultz and P.B. Armentrout, J. Phys. Chem. 91, 4433 (1987); R.H. Schultz, J.L. Elkind, and P.B. Armentrout, J. Am. Chem. Soc. 110, 411 (1988); R.H. Schultz and P.B. Armentrout, in progress. 18. K.K. Irikura and J.L. Beauchamp, J. Am. Chem. Soc. 113, 2769 (1991); T.J. McMahan, Y.A. Ranasinghe, and B.S. Freiser, J. Phys. Chem., submitted. 19. R. Tonkyn and J.C. Weisshaar, J. Am. Chem. Soc. 108, 7128 (1986). 20. S.W. Buckner and B.S. Freiser, J. Am. Chem. Soc. 109, 1247 (1987); S.W. Buckner, J.R. Gord, and B.S. Freiser, J. Am. Chem. Soc. 91, 7530 (1989). 21. C.E. Brown, S.A. Mitchell, and P.A. Hackett, J. Phys. Chem. 95, 1062 (1991) and references therein. 22. D. Ritter and J.C. Weisshaar, J. Am. Chem. Soc. 112, 6425 (1990); D. Ritter and J.C. Weisshaar, J. Phys. Chem. 94, 4907 (1990). 23. D. Ritter, Ph.D. Thesis, University of Wisconsin-Madison, 1990. 24. C.E. Brown, S.A. Mitchell, and P.A. Hackett, Chem. Phys. Lett. 191, 175 (1992). 25. P-O. Widmark, B.O. Roos, and P.E.M. Siegbahn, J. Am. Chem. Soc. 89, 2180 (1985). 26. J. Carroll, K. Haug, andJ.C.Weisshaar, work in progress. 27. C.W. Bauschlicher, Jr., Chapter in this book and references therein. 28. See, for example, M.R.A. Blomberg, P.E.M. Siegbahn, U. Nagashima, and J. Wennerberg, J. Am. Chem. Soc. 113, 424 (1991). RECEIVED January 21, 1993

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

Chapter 17

Gas-Phase Formation of Atoms, Clusters, and Ultrafine Particles in UV Laser Excitation of Metal Carbonyls Yu. E. Belyaev, Α. V. Dem'yanenko, and A. A. Puretzky

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch017

Institute of Spectroscopy, Russian Academy of Sciences, 142092 Troitsk, Moscow Region, Russia

Three main topics of laser-induced chemistry of organo-metallics are discussed - delayed luminescence of metal atoms, formation of electronically excited metal dimers and gas-phase production of ultrafine particles. Gas-phase U V multiphoton excitation (MPE) and dissociation (MPD) of organometallic molecules has been the subject of intense studies in recent years. This process has a few specific features which make it attractive for both fundamental studies and various applications. One main goal of fundamental studies is describing the mechanism of bare metal atoms production. MPD of organometallic molecules leads to the formation of metal atoms in ground and excited states (7-5) and also gives rise to a set of highly reactive coordinatively unsaturated species (6-8) which can initiate and participate in secondary chemical reactions. That means that MPD of such molecules can be used for basic studies on Laser Chemistry of Organometallics. In this chapter we shall discuss three closely related and we believe central topics of laser chemistry of Mo(CO)6The first topic is the formation of long lived highly excited Mo atoms. These energized atoms could start chemical reactions with the parent molecules and the coordinatively unsaturated compounds obtained in U V excitation of Mo(CO)6- It was found that under relatively high laser fluences (300 to 500 mJ/cm ) the so called "delayed" luminescence from high lying e Dj levels of Mo atoms was observed. The delayed luminescence pulse attains its maximum, about a few hundreds nanoseconds, after the end of the laser pulse and drops with a characteristic time of about a few microseconds, while spontaneous radiative lifetime of the observed transition z P°-e D is only a few nanoseconds. Detailed analysis has shown that this phenomenon can be explained by selective population of the e Dj levels through cascade transitions from the Rydberg states produced in U V MPE of the parent Mo(CO) gas. The second topic is related to the formation of electronically excited metal dimers. The attempts to produce ions of small metal clusters from stable organometallic compounds, such as Mk(CO) , (M is a metal atom, k=2-4, n=8, 2

7

7

7

7

6

n

0097-6156/93/0530-0220$08.50/0 © 1993 American Chemical Society

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

17.

BELYAEVETAL.

UV Laser Excitation of Metal Carbonyls

221

10, 12) by eliminating ligands with U V laser radiation are known well (see Review (9)). When such organometallic compounds are photoionized by U V laser, the mass spectrum has, along with a metal ion, dimeric, trimeric and tetrameric peaks of Μ 2 ' M3+» M4+ (9). We have observed neutral electronically excited dimers M02 upon excitation of Mo(CO)6 gas with XeCl laser (10) The parent compounds for them are the intermediates Mo2(CO)i formed from the Mo(CO)6 by the first XeCl laser pulse. The second pulse of XeCl laser removed the ligands form Mo(CO) 1 and gave rise to M02 dimers in the electronically +

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch017

ι

excited state Α Σ^. These dimers were detected by luminescence. Electronic absorption spectra of M02 were comprehensively studied in (11,12). Apart from such a laser-induced process, we have found out that electronically excited dimers, M02, are formed in a collisional chemiluminescent chemical reaction. The possible mechanisms of formation of electronically excited dimers M02 are discussed. The third topic we are discussing in this chapter is the formation of ultrafine particles in U V MPD of Mo(CO)6, Cr(CO)6 and W(CO)6 molecules in the gas phase. This process produces black soot which ICRFT mass spectrometer analysis has shown contains different molybdenum-carbon clusters. It has been shown that for Mo(CO)6 reactant, bright emission is observed under excitation of ultrafine particles by the second XeCl-laser pulse. The main characteristics of this emission are given. Experimental The experimental setup comprised two XeCl lasers (λ =308 nm, pulse energies Ei=30mJ, E =10mJ, pulse widths ( ^ = 15 ns (FWHM), τ = 10 ns (FWHM)) and a stainless-steel chamber (350 mm χ 350 mm χ 120 mm) having quartz or NaCl windows ( 0 60 mm). For two-pulse experiments, the laser beams were made to counter-propagate and focused into the chamber with quartz lenses (f 1 =20 cm, f2=25 cm, 0 60 mm). The cross-sectional areas of the both beams in the observation region were 2.5x8 m m and 3.5x5 mm , respectively. The pulsed C O 2 laser used for studying delayed atomic luminescence produced 2 J pulses having a 100ns (FWHM) peak and 500 ns (FWHM) tail. 2

2

2

2

5

The chamber was evacuated to a residual pressure of about 10~ Torr by means of an oil-diffusion pump fitted with a nitrogen trap. The container mounted directly on the chamber was filled with crystalline Mo(CO)6- The Mo(CO)6 pressure range used was 0.02 to 0.1 Torr. Mo(CO)6 was commercialgrade and was sublimed and stored under vacuum before being used. The luminescence spectra of dimers were analyzed by means of a monochromator (slit spectral width 20 Λ/mm) and a gated optical multichannel analyzer (OMA) (the spectral response of this system was 350 to 850 nm, spectral resolution 2 Â, minimum gate width 150ns). The temporal evolution of luminescence at certain fixed wavelengths was studied using a photomultiplier (spectral response 200 to 800 nm). To avoid the influence of the diffusive escape of particles from the observation zone in the time range 0-10 μ s on the measured kinetics we had to choose a transverse observation geometry. The elongated exposed zone was projected (scale 1:1) by a lens (f=10 cm) not along the monochromator slit (height of the slit h=1.0 cm) but across it. A small region of this zone was registered. Our calculations have shown that with such a irradiation and detection geometry the expansion of the exposed zone even due to free particle flying apart does not

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

222

LASER CHEMISTRY O F ORGANOMETALLICS

affect considerably the measured time dependencies with ι At . In the pressure range max

studied

max

is proportional to pL8±0.3 f

or

At ο and to p0.8±0.2 f

o r

Ar

m a x

.

Let's consider the processes which may result in the observed dependences. The first laser removes ligands and gives rise to a set of fragments Fk=Mo(CO)k where 0< k F ,

(5)

k

where M=Mo(CO)6 denotes parent molecules. Then, the Fk fragments reacting either with the parent molecules M or with the slightly stripped fragments Fk, form molecular dimers Di=Mo2(CO)i. These dimers reacting with M or Fk, with the rate constant k2 produce a certain product Ρ F +M(F ,)-±->D M{F )-^P k

k

l+

k

lnh(û{2) D

u

D, 0

D* , 0

\

(6) />"...

I

A

»L

The second laser pulse coming with a variable delay, Δί, removes the ligands from the molecular dimer Dj thus producing either highly stripped molecular dimers D i , or metal dimers D =Mo2 distributed over electronically excited states. By monitoring the luminescence of M02 from a definite electronically excited state we can follow the kinetics of formation of molecular dimer, D i . 0

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

BELYAEV ET AL.

UV Laser Excitation of Metal Carbonyls

V'

0 I

I

2 1

3 I

4 I

5 1

6 I

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch017

ιI w

514

516

518 520 522 524 WAVELENGTH,nm

526

b

W A V E L E N G T H , nm

Figure 10a, b. M02 luminescence spectra for the peak (a) and the tail (b) of the luminescence pulse. (Adapted from ref. 10)

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

237

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch017

Figure 11.

Spectrum integrated M02 luminescence intensity vs time delay between XeCl-laser pulses. (Reproduced with the permission from ref. 10. Copyright 1990.)

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

17. B E L Y A E V E T A L .

239

UV Laser Excitation of Metal Carbonyls

The formation and disappearance of molecular dimers D i is described by simple rate equations. As a result, we can derive from them an expression for the concentration, D i [Di ] = [ F * k -r^r

· {exp(-*,[Af]r) - exp(-* [M]r)}, 2

(7)

when [FiJo is the initial concentration of Fk fragments. The curve obtained from Eq. (7) is shown in Figure 11 by a solid line. From comparison of experiment and calculation for the constants k i is in a good agreement with the rate constants for the reactions of Mo(CO)4 and Mo(CO)3 with the parent molecules Mo(CO)6 given in (77) (Table I). With the laser fluences varying between 100 and 200 mJ/cm the probability of induced single-photon absorption is σΦγ= 3. Therefore, the volume under irradiation contains a high concentration of relatively long-lived slightly stripped fragments Fk, with k'=5. It is these fragments that along with the parent molecules M are obviously the most probable partners for Fk in reaction (6). The integral intensity of luminescence is proportional to [Di] which shows that with A i , 7 j ^ is proportional to p. When At is small, it follows from (7) that

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch017

2

m a x

The both cases are in a good agreement with the experimental dependences (10). Luminescence pulse tail. The luminescence pulse tail of M02 dimers has some specific features which are not described by the mechanism for the peak discussed above. First of all, this is relatively long luminescence. It should be ]

noted that the radiative time of luminescence for the state ΰ' = 0, Α Σ ^ is just 18 ns (72). Figure 10b shows a luminescence spectrum of M02 dimers measured for the luminescence pulse tail. The time parameters in this case were: the delay between the laser pulses At = l , the gate duration Ag = 500ns, the delay between the onset of the second laser pulse and the gate At = 700ns. This spectrum differs greatly from the similar spectrum measured for the peak of M02 dimers by the intensity distribution of the related lines. It should be also noted ι

that the population of the vibrational level = 0,Α Σ^ is lower compared to the vibrational level ϋ' = 1. The characteristic feature of the luminescence spectrum is retained when the delay At is varied. It has been shown that l( ) is proportional to p^iO.S^ j pressure dependence is stronger for the pulse tail than for the peak. This explains the qualitative effect of disappearance of the pulse tail as the Mo(CO)6 pressure decreases. The experimental dependences of the luminescence intensity on different experimental parameters can be explained by the following mechanism. Some D

e

m

e

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

240

LASER CHEMISTRY O F ORGANOMETALLICS

dimers, M02, (below referred to as active dimers

j) produced by the second

r

laser pulse react with the parent gas Mo(CO)6 ° with the fragment Fk, forming x

highly excited unstable complexes, like T - Mo^CO)*^ which dissociate to form n

electronically excited dimers M02 in the state A j Z ^ (û

k

D j+M(F^ ) — - ^ T n — ^ D * + A + nL—^D +hœ 0

0

(9)

L

Here, A is the M o atoms, L is the CO ligands and Ks=l/r . Under the assumption that k4 » k3, k5 the sequence of the processes (9) is described by rate equations from which it follows that July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch017

rad

D

I ~k [D* }^[D* }'m cxp{-k [M]t) 5

0

0

r

(10)

3

Here, t - Δτ. It is assumed in (10) that k5» k3[M]. From (10) it follows that with 2

small delays At between the laser pulses, when D ^ ~ p and At= At < , when m

I

{

D

)

àx

~ P \ P \

(ID

if At is small enough. Single-pulse Excitation. In the case of MPE by one XeCl-laser pulse with fluences 300 -500 mJ/cm one can observe a luminescence band corresponding to 2

the transition Χ Σ^>-

LL

/ / \ \

ζ LU h-

\

/

* 3

DELAY

Figure 17.

-J

I I I

J ί m

* ι

TIME,

D

ι ο

1

>

/ /

τ

m or
.04, but for very high power and slow scan speeds, the coalescence becomes an unstable process, leading to gaps in the gold film that is formed. The residue of gold that remains near the gaps

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

20.

COMITA E T A L .

Laser-Induced Coalescence of Gold Clusters

Ί

1

1

1

1

Γ

10"

Λ"· E ΙΟ"

4

/ *

ο

/

G ΊΟ >

I

I B u l k Resistivity of G o l d : 2.35 1 0 Ω-cm

A

- 6

-2

10°

00

ÛC ο

2

10 10

4

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch020

•ft ft 0.0

0.2

0.4

0.6

0.8

L 1.2

1.0

1.4

Metal V o l u m e F r a c t i o n

Figure 4. D C resistivity of the as deposited films as a function of gold volume fraction.

Table I. Resistivity of Laser Annealed A u / P P F C Films Au

Film

Power

Scan

Line

Resisitivity

Volume

Thickness

(mW)

Speed

Thickness

(/iohm-cm)

Fraction

(A)

(Ws)

(nm)

0

14000

4

23000

20

100

500

9.1x105

16

10000

50

10

160

3.66

21

1700

148

200

40

4.6x104

20

7000

104

10

100

2.70

23

6000

198

10

200

2.66

24

16000

190

200

800

4.83

28

13000

40

20

820

26.4

30

12000

40

20

720

97.6

1016

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

285

LASER CHEMISTRY OF ORGANOMETALLICS

. I . I . I . III

. ι

ι

ι , ι .

0

(b)

1

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch020

0.4

0.8

(c)

1.2

-

1.6

Γ

(d)

-

2.0

2.4 2.5

(e) -

III , 1 . 1 • 1 . ~ 4 4 20 36 52 0 Radial Dimension (microns)

I

52

-

I - •

36 20

Figure 5. Surface profile of laser-annealed film showing the extent of polymer removal and film collapse with increasing laser power and constant scan velocity. The scan velocity is 100 μ m/s and the incident laser power (mW) is (a) 5.8; (b) 11.6; (c) 29.4; (d) 58.8; and (c) 90.9.

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

COMITA ET AL.

Laser-Induced Coalescence of Gold Clusters

287

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch020

20.

Figure 6. A n S E M photograph of the laser annealed A u / P P F C . The periodic features are due to the morphology of the film. In this film, the ρ = .24 with an as deposited film thickness of 16000 angstroms. The laser power was 76 mW at 514 nm. In the upper photograph, the linewidth is approximately 10 microns. The lower photograph is at 20,000 X magnification, and shows the continuous nature of the film.

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch020

288

LASER CHEMISTRY OF ORGANOMETALLICS

0.0

1.0

2.0 X-ray

3.0 Energy

4.0

5.0

(keV)

Figure 7. Energy dispersive x-ray analysis of the laser-annealed film. The major line is gold, with trace amounts of other elements, including fluorine.

Table II. Surface Temperature Rise during Laser Coalescence Power (mW)

Radius (/im)

Depth (/im)

Temp. Rise (C)

5.8

4.71

0.1

70.6

11.6

7.28

0.4

91.3

29.4

8.57

1.1

196.6

58.8

13.5

1.8

249.6

90.9

18.8

2.4

277.0

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

20.

COMITA ET AL.

Laser-Induced Coalescence of Gold Clusters

289

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch020

are generally spherically shaped nodules, indicative of complete melting of the gold film before a continuous film has a chance to form. We would like to estimate the maximum surface temperature induced by the laser in order to understand the physical and chemical mechanisms which are possible in the estimated temperature range. For the purpose of making an estimate of the temperature rise, we have assumed that the polymer undergoes simple volatilization, and the gold clusters migrate and coalesce to form a continuous film. We can estimate that with gold clusters of approximately 50Â average diameter and ρ = 4%, that after removal of 125 nm of film we have an effective p = 100% in the top surface layer, approximately 5 nm in depth. To a first approximation, this layer will absorb the visible light and give rise to a heated region which then volatilizes the film beneath it causing further collapse and additional gold coalescence. The beam diameter is much larger than the thickness of the gold layer, so we can neglect thermal losses through the gold layer, and use the thcrmophysical properties of the film/substrate. The beam is also larger in diameter than the thickness of the A u / P P F C film, and to a first approximation the thermal conductivity of the S i 0 can be used as the surrounding heat sink. 2

The temperature rise can be calculated using the steady state expression (12): T = P(l - R ) / 7 r r K . In this expression Ρ is the incident power, R = 0.64, the reflectivity of A u at 514 nm (13) r is the radius of the gold film at the coalescence zone, and K =0.02 W/cmC, the thermal conductivity of the quartz substrate. Using the range of laser fluence that we generate conducting lines with (see Table II), we can estimate that the temperature rise extends from 70C to 277C. The temperature present in the film under irradiation could be somewhat higher than that indicated for a number of reasons which we will enumerate here: 1) the reflectivity may be much lower because the films are not continuous and smooth as they are being formed; 2) the teflon film has a lower thermal conductivity than the substrate, and this will lead to less heat loss; 3) the annealing process is highly exothermic due to the polymer and the gold clusters being initially formed in a metastable state in the plasma deposition process. The additional energy input due to the relaxation phenomena occurring in the matrix can be substantial and we have not taken this into account. We may regard the number calculated as being a lower estimate of the surface temperature maximum actually achieved during the laser annncaling process. However, they give an indication of the approximate temperature range. m a x

f

s

f

s

In order to gain an understanding of the physical and chemical phenomena involved in the laser-induced coalescence, we have examined the optical properties of the system by using a time-rcsolvcd reflectance experiment. In this experiment we measure the specular reflectance at a fixed wavelength during a single coalescence event in the thin film. This is accomplished with an experimental system which consists of an A r + laser as the light source, the output of which is modulated by a 25 M H z Κ D P pockels cell. A n experimental set-up very similar to this has been described previously (14). The specular reflectance was collected with a high numerical

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

290

LASER CHEMISTRY OF ORGANOMETALLICS

aperture lens, dispersed in a monochromator, and detected with a photomultiplicr tube. The reflectance signal was captured with a transient digitizer interfaced to an I B M PS/2. The change in optical properties is observed to occur under bulk annealing at approximately 200C, very close to the glass transition (T ) of P P F C . The optical change is correlated with the coalescence of the metal clusters, and can be described as taking place with the collapse of the polymer phase and migration of small clusters to form larger metal clusters. It has been determined previously that the size of the clusters (30-50 A ) can give rise to a strong change in transmittancc in the visible region of the spectrum, and this is expected to induce a change in the reflectance of the sample as well, through a modification of the optical constants of the film. This effect is being further examined to quantify the change in the optical properties.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch020

g

The time-resolved data was obtained for the change in specularly reflected light for thin film samples with moderate gold loadings. Again, these are films which are below the percolation threshold, so that they are useful as a dielectric film or matrix for the resulting conducting patterns. We have found that the reflectance change can be fit to a exponential function, and from this fit we are able to calculate a a half life of 455 //sec (at 10 μηι, and 150 mw). The time scale for complete collapse of the composite film to a continuous gold film can be seen in Table I, for a Κ) μνη beam. The beam velocity is 100 - 200 μηι/sec at the fastest scan rates. The complete physical collapse to a continuous metal line thus takes place on the order of 100 msec. This indicates that the large reflectance changes occur in the initial stages of coalescence before the film completely collapses. The approximate time for spheroidization of two gold spheres can approximately be calculated by a previously described analytical expression (15). This treatment uses the surface energies as well as the diffusive flux of metal atoms to roughly predict the time of sintering and the time of complete spheroidization of two contacting particles. This can be calculated to be approximately 1 ^scc at 150 C for a 40 À cluster size with a relatively straightforward model (15). This time is below that which has been measured for the reflectance change, but is more closely approximated by a particle with a larger radius. We expect that the rate limitation is therefore cither coalescence of large clusters, which have already coalesced from their initial state, or simply migration of the clusters together, by the pyrolysis or volatilization of the polymer film. Summary and Conclusions In this work, it is shown that bulk heating or annealing of the film leads to film collapse from a metastable polymer state, with concomitant microstructural changes in the size and shape of the gold clusters. Laser-induced coalescence of the gold clusters gives rise to metallic lines with approximately the linewidth of the 1/e intensity points of the laser beam. The extent of film collapse can be controlled be both scanning speed or dwell

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

20. COMITA ET AL.

Laser-Induced Coalescence of Gold Clusters

291

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch020

time as well as the laser power. A t high gold loadings (>4%) the film collapses with a quasi-Gaussian profile, similar in shape to the laser intensity. The depth of the collapse is dependent on the dwell time and laser intensity, but at very low powers collapse is complete for most films. The polymeric matrix undergoes simple volatilization, and the gold clusters coalesce to form a continuous film. To a first approximation, the metal absorbs the visible light and gives rise to a heated region in the top surface which then volatilizes the film beneath it causing further collapse and additional gold coalescence. A t the very lowest powers the resistance is far above that for a bulk gold sample, and as the power increases, the resistivity approaches that of bulk gold. The technological significance of this work lies in the ability to directly generate microcircuit patterns in and on a teflon-like carrier. The method is far simpler than those recently described in that the films are formed in a one step dry process, and the metal line patterns are then directly written in a single subsequent step. The patterning step can take place in an ambient air environment. Gold conductors are equivalent in conductivity to copper, but arc more robust to oxidation, and thus do not need passivation. In addition, the pure polymer films can be subtractivcly patterned with a single step laser exposure. The films can thus be built up into a 3-dimensional structure with metal lines and/or vias, making this an extremely versatile system for microelectronic circuitry or packaging applications.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

Price, P. E.; Jensen, K. F. Chem. Eng. Sci. 1989 44 1879-1891. Fisanick, G. J.; Gross, M. E.; Hopkins, J. B.; Fennell, M. D.; Schnoes, K. J.; Katzir, A. J. Appl. Phys., 1985 57: 1139-1142. Marcus, H. L.; Beaman, J. J.; Barlow, J. W.; Bourell, D. L. Amer. Ceramic Soc. Bull., 1990 69: 1030-1031. Perrin, J.; Despax, B.; Hanchett, V.; Kay, E. J. Vac. Sci. Technol. A 1986, 4, 46. Laurent, C.; Kay, E.; Souag, N. J. Appl. Phys. 1988, 64, 336. Kay, E.; Hecq, M. J. Appl. Phys. 1984, 55, 370. Kay, Ε. Z. Phys. D 1986, 3, 251. Kay, E.; Dilks, A. J. Vac. Sci. Technol. 1981, 18, 1. Laurent, C.; Kay, E. J. Appl. Phys. 1989, 65, 1717. Perrin, J.; Despax, B.; Kay, E. Phys. Rev. Β 1985, 32, 719. Comita, P. B.; Zhang, R.; Jacob, W.; Kay, E., unpublished work. Carslaw, H. S.; Jaeger, J. C. Conduction of Heat in Solids 2nd Ed., Oxford University Press, Oxford, 1959 p 264. Palek, E. D. ed. Handbook of Optical Constants of Solids, Academic Press, Orlando, 1985. Comita, P. B.; Price, P. E.; Kodas, T. T. J. Appl. Phys. 1992, 71 , 221. Blachere, J. R.; Sedehi, Α.; Meiksin, Ζ. H. J. Materials Sci. 1984, 19 , 1202.

RECEIVED February 10, 1993

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

Chapter 21

Laser-Assisted Chemical Vapor Deposition from the Metal Hexacarbonyls 1

K. A. Singmaster and F. A. Houle

2

1

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch021

2

Department of Chemistry, San Jose State University, San Jose, CA 95192-0101 IBM Research Division, Almaden Research Center, 650 Harry Road, San Jose, CA 95120

Laser induced deposition of thin films from the group VI hexacarbonyls can be accomplished by either photolysis of the gas phase precursor (photochemical deposition) or by laser heating of a substrate in the presence of the organometallic vapor (thermal deposition). Towards elucidating the microscopic processes relevant to photochemical and thermal deposition, a systematic study of compositions of films deposited by cw-257 nm and cw-514 nm irradiation have been performed. UV photolysis is known to lead to partial decarbonylation of the hexacarbonyl; additional surface photodissociation and dissociative chemisorption of CO produce films which have high levels of C and Ο impurities. On the other hand laser heating of a substrate in the presence of the vapor with visible light produces significantly cleaner metal films. Under these conditions dissociative chemisorption of CO is not observed. Thermal modeling and the relationship between this work and low coverage surface studies will also be discussed.

The technique of laser assisted chemical vapor deposition ( L C V D ) offers great promise for localized film growth at microscopic scale (1). In L C V D a substrate held in a cell backfilled with metallorganic vapor is exposed to a laser source; metal containing material is deposited onto the surface of the substrate. Film deposition is the result of photochemical and/or thermal processes. Choosing a laser wavelength that corresponds to light absorption by the metallorganic vapor and/or surface absorbed species can result in photochemical reactions which lead to film deposition. If, on the other hand, the substrate absorbs the N O T E : This chapter was originally presented at the fall 1991 meeting of the Electrochemical Society, Inc., held in Pheonix, AZ.

0097-6156/93/0530-0292S06.00/0 © 1993 American Chemical Society

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch021

21.

SINGMASTER AND H O U L E

LCVDfromthe Metal Hexacarbonyls

293

laser energy, thermal reactions of the precursor on the substrate surface can produce metal containing films. Separation of these two mechanisms of deposition is not always possible and requires careful choice of precursor, substrate and laser conditions. The group V I metal hexacarbonyls have the potential to be excellent precursors for film deposition by ultraviolet photolysis; they are reasonably volatile, have high cross sections in the U V and dissociate in the gas phase with near unity quantum yield to form partially carbonylated species (2-6). By using low power U V laser sources and choosing an appropriate substrate it is possible to accomplish photochemical deposition in the absence of laser heating. O n the other hand, the hexacarbonyls are transparent to visible light thus by using a visible laser and a substrate that absorbs visible light it is possible to observe thermal deposition in the absence of photochemical processes. Films deposited by photochemical deposition with low power U V lasers are of poor quality, containing large levels of carbon and oxygen impurities (5-7). High levels of impurities produce large resistivities making these films useless to the microelectronics industry. Cleaner films can be obtained with higher power densities (typically obtained with pulsed laser sources) which can produce significant surface heating (7,8). Few studies have been performed which investigate the composition of films thermally deposited from the metal hexacabonyls. Studies with W ( C O ) indicate that clean metal films can be deposited on 2 micron thick Si substrates (514 nm source), while only 75% metallic deposits were obtained on quartz (9,10). The decrease in impurities observed for thermally deposited films as compared to photochemically deposited films is qualitatively consistent with the known surface chemistry of C O on W single crystals, where surface temperatures greater than 1000 Κ result in recombination of C and Ο and desorption of C O from the surface (11). This paper outlines systematic composition studies we have performed on both photochemical and thermal deposition from the group V I hexacarbonyls. The goal of this work is three fold: to provide a complete composition data set for these films under controlled vacuum and analysis conditions, to provide insight into the methods by which contaminants are incorporated into these films, and to contrast the results obtained to previous surface and gas phase studies of metal hexacarbonyls. 6

Experimental Conditions A l l films were deposited on Si(100) or (111) doped n- or p-type, cleaned in acid oxidant bath and left covered by its native oxide. The light source for the laser heating work was an A r ion laser focused to a 10 micron spot; for the photochemical work the frequency was doubled and the U V irradiation focused to a 5 micron spot. Under these conditions incident power densities ranging from 40 to 3,700 W / c m were available at 257 nm ( U V ) while 1 - 4 M W / c m was utilized for the visible irradiation (514 nm). Two deposition cells were used: 2

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

2

294

LASER CHEMISTRY OF ORGANOMETALLICS

4

a stand alone cell which could be evacuated to the mid 10" Torr range, and an ultrahigh vacuum cell with a base pressure in the 10" Torr range. The latter could be coupled to a vacuum suitcase for transportation to the analysis chamber without air exposure. For all depositions the cells were backfilled with the vapor of crystalline M ( C O ) ( M = Cr, M o or W) which had been subjected to several freeze-pump-thaw cycles. No buffer gases were used. Room temperature vapor pressures for the three hexacarbonyls are; 125 mTorr for Cr, 85 mTorr for M o and 17 mTorr for W. Absorption coefficients at 257 nm were measured for all three materials (12). Three sets of measurements were carried out for each metal for both photochemical and thermal deposition conditions: growth in low vacuum and exposure to air prior to analysis (LV/air), growth in high vacuum and exposure to air ( H V / a i r ) and growth in high vacuum with transfer in vacuum for analysis (HV/vac). A l l films were characterized with a scanning Auger multiprobe using a 10 keV electron beam. Depth profiles are obtained with a 2 keV A r ion sputtering beam (6,2). Calibration of beam damage and preferential sputtering effects was accomplished by using standard oxides and are outlined in Ref. 6. The beam conditions chosen significantly hinder all but instantaneous beam damage effects. C r 0 was the only oxide that did not undergo preferential sputtering of oxygen under the conditions of analysis, thus O / M ratios for sputtered W and M o films only provide a lower limit to the oxygen content. 9

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch021

6

2

3

Results Tables are included for both photochemical and thermal deposition. Only H V / v a c data is tabulated. For L V / a i r and H V / a i r data see Ref. 6 and 12. Each entry is an average of compositions from 12 - 50 films. Peak-to-peak height ratios of Ο ( K L L , 503 eV) to Cr ( L M M , 529 eV), Ο to M o ( M N N , 186 eV), Ο to W ( M N N , 1736 eV), C ( K L L , 272 eV) to Cr, C to M o and C to W are reported together with the spread in measurements. The percent compositions correspond to the average values and are provided for purpose of comparison to previous studies. Data obtained for authentic oxide and carbide samples and predicted peak height ratios for selected stoichiometries are also included at the end of each table. Comparison of predicted ratios with the experimental data permits assessment of approximate stoichiometries for the films. Photochemical Deposition. Photochemically deposited films exhibit deep ripples characteristic of deposition with a linearly polarized laser beam under conditions where photo-induced surface reactions are rate limiting (12,13). This permits comparison of surface and gas phase photochemical and photophysical processes. Ripples were observed at all power densities used in the photochemical studies. Deposition rates ranged from 3,000 A / s to 10 A / s depending on the precursor and the power density (12). Table I, II and III contain the Auger data for Cr, M o and W H V / v a c films deposited from the corresponding hexacarbonyl by photolysis with 257 nm

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

21.

SINGMASTER AND H O U L E

LCVD from the Metal Hexacarbonyls

295

light. Depth profiling showed all films to be homogeneous in composition after removal of the top 1 - 2 nm, so data for sputtered films were averaged over the profile. Data for unsputtered films reflect compositions of the near surface region. No significant variations in composition were found over the laser power density range used, so data for all intensities are combined. Because of extensive gas phase photolysis during deposition a thin (5 nm) metal containing film covers the substrate. Analysis of this off spot material is

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch021

Table I. Auger data for H V / v a c Cr films photochemically deposited O/Cr

C/Cr

Ο

C

Cr

FILMS unsputtered sputtered off spot

2.5+0.6 1.4+0.3 2.3+.0.1

0.23 + 0.07 0.23 + 0.07 0.43+0.03

54 39 43

19 25 33

28 36 24

C r 0 Standard unsputtered sputtered

2.4 + 0.1 2.0±0.1

2

3

Predicted Cr 0 CrCO CrC 0 2

3

2

2

1.9 1.2 2.4

35 39

65 61

0.28 0.57

60 33 40

33 40

40 33 40

Table II. Auger data for H V / v a c M o films photochemically deposited O/Mo

C/Mo

Ο

C

FILMS unsputtered sputtered off spot

1.5+0.4 0.8+0.2 1.5+0.3

0.63.+0.10 0.59+0.10 0.73+0.15

27 16 24

46 53 52

M o 0 Standard unsputtered sputtered

5.7+0.4 3.4+0.4

Mo 27 30 24

3

21 31

79 69

Predicted M0O3

MoCO

MoCA

4.5 1.5 3.0

0.34 0.68

75 33 40

33 40

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

25 33 40

296

LASER CHEMISTRY OF ORGANOMETALLICS

Table III. Auger data for H V / v a c W films photochemically deposited

o/w

c/w

Ο

C

W

FILMS unsputtered sputtered off spot

1.8+0.4 0.9+0.2 1.8+0.3

0.7+0.2 0.5+0.1 0.7+0.2

20 12 20

33 30 33

47 58 47

W 0 Standard unsputtered sputtered

6.4 + 0.5 3.8jt0.4

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch021

3

W C Standard unsputtered sputtered

40 57

60 43

1.0+0.1 1.2+0.2

50 55

50 45

33 40

25 33 40

Predicted

wo wco wc o 3

2

2

12.8 4.3 3.0

75 33 40

0.98 0.68

listed in all three tables. Comparison of H V / v a c and H V / v a c off spot films probes differences in surface reactions of M ( C O ) with and without incident U V photons. No significant differences are found in M o and W systems, but the Cr off spot films show roughly double the amount of carbon and oxygen. A l l carbon line shapes were either graphitic or carbidic, indicating that any C O groups remaining in the films are fully dissociated. Auger data for L V / a i r and H V / a i r (reported elsewhere) show that film compositions depend strongly on deposition conditions (6). The L V / a i r films have as much oxygen in them as the thermodynamically stable oxide. Comparison of L V / a i r and H V / a i r films show that oxygen containing background gases react with the films during deposition decreasing the amount of carbon in them. A i r exposure of films deposited under high vacuum conditions (HV/air) results in much higher oxygen content than H V / v a c films. In addition, the carbon distribution becomes inhomogeneous, with the bulk concentration being much lower than that of the surface. x

Thermal Deposition. Film morphologies and topographies varied widely for thermally deposited films (14). Rippled film structures were not observed. Due to the varied topographies only an estimate of deposition rates for M o were measured (2,000 - 6,000 A / s ) . Table I V incorporates all the H V / v a c data for Cr, M o and W films thermally deposited with 514 nm light. The compositions are spatially inhomogeneous so data for film centers and edges are included. Film

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch021

21.

SINGMASTER AND H O U L E

297

LCVDfromthe Metal Hexacarbonyls

compositions remained constant after the first 1 - 2 nm of film were sputtered off, thus sputtered data includes a variety of depth profile data. The data for centers and edges exhibited no variation in composition over laser power density used, so data for all intensities are combined. Metal-containing material was observed on the substrates in regions far from the films as well as in between films. The thickness of this deposit depends on the distance from the spot and the irradiation time of the spot, and was sometimes thick enough to be sputtered (> 10 nm). Sputtering was only performed for M o . Although oxides of W and M o are well known to undergo preferential sputtering of oxygen, it is highly unlikely that this occurs in the thermally deposited films. Both M o and W centers exhibit drastic reductions in oxygen content within the first 2 nm. Typical W and M o oxides lose some but not all oxygen within the first few nm. Clean film centers were observed for sputtered material (HV/vac); this composition was not significantly affected by a five minute air exposure (HV/air). For M o and W the presence of background gases during deposition ( L V / a i r ) only added relatively small levels of oxygen to the films while in the case of Cr the well documented formation of C r 0 was observed. The lack of significant oxygen incorporation during air exposure of H V / a i r films is attributed to the density of the films. Differences between edge and center data arise only during sputtering; the surface compositions are uniform indicating that an additional layer of gas phase material decomposes on the film surface after the irradiation is terminated. 2

3

Table IV. Auger data for thermally deposited H V / v a c films Film

M

Sputtered C/M

C r Films center 0.23 + 0.15 1.6 + 0.8 edge 0.23.+ 0.10 1.6 + 0.4

42 26 32 42 26 32

O.OjfO.O 0.08 + 0.05

M o Films center 0.46 + 0.15 1.2+0.4 edge 0.58 + 0.30 1.4 + 0.4 off spot 0.97 + 0.25 2.1 + 0.4

25 43 32 0.0+0.0 0.0+0.0 0* 0* 100 26 47 28 0.10±0.10 0.2+0.2 8 21 77 27 54 19 0.32 + 0.05 0Λ+0.1 12 43 71

W Films center 0.44 + 0.20 1.8.+ 0.8 edge 0.47 + 0.15 1.8 + 0.8

22 24 53 22 25 53

C/M

Unsputtered Ο/Μ O

C

0.0+0.0 0.0+0.0

O/M

Q

C

M

0.3+.0.3 19 0* 81 0.8 + 0.3 33 15 52

0.0+0.0 0* 0* 100 0.0+0.0 0* 0* 100

- below detection limit of Auger

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

298

LASER CHEMISTRY OF ORGANOMETALLICS

It is important to note that oxygen-free film centers for Cr films were observed on some occasions. The data on the table represent averages over many films. A n Ην/vac run using a power density of 0.7 M W / c m at 257 nm was performed to determine the effect simultaneous laser heating had on the composition of photochemically deposited film. This power density represented a 200 fold increase from that used in the photochemical work. Unfortunately, window deposits significantly and rapidly attenuated the beam so the extent of heating was most likely significantly less than one would hope. Significant differences between high power U V and low power U V compositions were not observed. Additional discussion regarding these studies can be found in reference 14. July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch021

2

Discussion Photochemical Deposition. Although exposure to background gases during and after deposition significantly alter the contamination levels of the films and are of great relevance to possible applications, this discussion will be limited to H V / v a c films. It is these films that provide insight into the deposition process and possible clues to methods that might result in lower impurity levels. It has been suggested that film growth from the metal hexacarbonyls is mediated by gas phase photolysis, that is, surface reactions of M ( C O ) and M ( C O ) are responsible for film growth at 257 nm. Under the present conditions, gas phase reactions are saturated and surface photochemical reactions are rate limiting. Examination of H V / v a c film compositions suggests that the film stoichiometries are consistent with CrCO, M o C 0 and W C 0 . Away from the laser beam (off spot) stoichiometries are consistent with C r C 0 , M o C 0 and W C 0 . Thus it appears that condensation of M ( C O ) and M ( C O ) leads to loss of about 2 to 3 C O groups (or C O and C 0 for Mo) via a thermal route. The remaining C and Ο cannot be removed photochemically for the M o and W system, as judged by film compositions in illuminated areas. In contrast, there exists a well-defined photochemical route for loss of one additional C O from Cr. It should be noted that even though the thermal and photochemical pathways are not chemically distinct in the M o and W systems, the photochemical reaction rate is much faster. It is interesting to compare these results to studies of surface photolysis of M ( C O ) condensed on single crystal Si, M o or R h surfaces at low temperatures (15). Although details of experimental procedures varied, the data consistently showed that decarbonylation under U V light is incomplete and that the C O groups remaining on the surface were not dissociated. Raising the temperature to 300 Κ resulted in the loss of additional C O . Dissociated C O still remained on the surface. In the present study, carried out under film growth conditions rather than monolayer adsorption, the initial adsorbate is already decarbonylated, the surface temperature is near ambient, and the surface is an amorphous metal oxycarbide. Despite the marked differences in reaction conditions, it appears that at least qualitatively the increased extent of C O loss 4

3

2

< 1

< 1

2

2

< 1

3

< 1

4

2

6

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

2

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch021

21.

SINGMASTER AND H O U L E

299

LCVD from the Metal Hexacarbonyls

and complete dissociation of remaining C O groups may be accounted for by the higher surface temperature. The surface composition may also affect the chemistry since preadsorbed C and/or Ο is known to inhibit dissociation of M o and C r and may thus facilitate C O desorption during deposition (16,17). One of the key differences between the film and surface studies is that the metal hexacarbonyl photoproducts arriving at the surface of the growing film can undergo subsequent photodissociation reactions, whereas primary photoproducts formed directly on single crystal surfaces at reduced temperatures do not. In addition the extent of C and Ο incorporation into the films is much higher than what would have been expected from coadsorption studies of C O / M o and C O / W (6). Contamination of photochemically deposited Cr, M o and W films by incomplete removal of CO, reaction with background gas during deposition and reaction with air is so efficient that we were unable to find reaction conditions under which pure metal films were formed in the absence of heating. Thermal Deposition. It is evident that the elemental compositions of H V / v a c films are not uniform across the film diameter; centers contain clean metal while the edges have C and/or Ο impurities. A l l three hexacarbonyls are transparent to 514 nm light; gas phase photochemistry does not contribute in any way to the film deposition process. Film deposition is the result of laser heating of the S i / S i 0 substrate and the growing film with subsequent thermal decomposition of the hexacarbonyl. Comparison to chemical vapor deposition work, although useful, must be done carefully since these depositions occur in the presence of carrier gas. These gases typically contain trace impurities and in some cases are chemically active. Kaplan and d'Heurle performed C V D studies on W ( C O ) and M o ( C O ) in the temperature range between 570 Κ and 820 Κ (18). Very low levels of carbon impurities were observed at Τ > 770 Κ. The carbon content increased significantly at lower temperatures. Oxygen content was not investigated. Others have also reported appreciable levels of C and Ο impurities incorporated into C V D films prepared at Τ < 770 Κ from M o and W hexacarbonyls (19). The C V D work suggests that in our M o and W laser experiments temperatures greater than 770 Κ might be obtained at the center of the M o films while the edges of the film and off spot material are below 770 K . Since the films are larger than the focal diameter of the laser it is not surprising to observe increasing degrees of contamination as a function of distance from the film center, representing a radial decrease in temperature. 2

6

6

In order to correlate composition and local temperature of the deposited films and to compare the laser deposition and C V D studies, radial temperature profiles for conditions corresponding to steady state film growth have been calculated. A stochastic method was used which incorporated both thermal and optical parameters for the films. The algorithm was tested with both deterministic approaches and experimental data. Additional details on this approach can be found in reference (14). Temperature profiles for a series of films were calculated. For W films deposited at 2.3 - 2.8 M W / c m , the calculated center temperatures are in the order of 900 to 1000 Κ while film 2

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

300

LASER CHEMISTRY OF ORGANOMETALLICS

edges reached temperatures in the 500 Κ regime. Since both edge and center data for W show no detectable impurities, these profiles indicate that temperatures around 500 Κ can produce clean W films. More interesting results are obtained for Mo. Direct comparison between profiles and film data indicate that impurities for a variety of laser power densities (1.5 - 4.0 M W / c m ) are not detected until temperatures below 400 Κ are reached. This is significantly lower than what was expected from the C V D studies. Similar results were obtained for incorporation of C impurities in Cr films. (Oxygen content in C r does not serve as a reliable marker for impurities since Cr is known to be a very efficient scavenger of oxygen.) Once local temperatures are available it is possible to compare the film results to low coverage surface studies of C O and hexacarbonyls on single crystal surfaces. Heating a W(110) surface saturated with partially carbonylated species and bound C O produced by the thermal decomposition of W ( C O ) results in both desorption and dissociation of C O (300 - 400 K). Complete removal of C and Ο from the W(110) surface is accomplished at around 1200 Κ (11,20). Studies of C O adsorbed on W and M o provide a range of temperatures over which dissociated C O recombines and is fully removed from the surface: for Mo(100) , Τ > 1200 Κ (16,21); for polycrystalline W, Τ > 1600 Κ (22); for W (100), Τ > 1300 Κ (23). Due to experimental difficulties very little is known about recombinative desorption of C O on Cr surfaces. Comparison to the films indicates that loss of all C O from the films is obtained well below any of the temperatures observed for surface studies. It is unlikely that recombinative desorption of C O is kinetically efficient in these films because the C and Ο concentrations are low. Thus it appears that all C O is lost through direct thermal desorption, that is, the dissociation channel is unimportant when films grow at temperatures higher than 400 Κ (24). This conclusion is supported by mechanistic simulations described in full elsewhere (14). This result is far from what surface scientists have suggested as the temperature range required to thermally produce clean films (well above 1000 K ) and serves to illustrate that direct film studies are essential towards the understanding of the mechanisms involved in deposition. Recently Houle and Yeh have performed mass spectroscopic studies on the laser thermal decomposition of Cr(CO) on Si during film growth (25). Their apparatus allows for the observation of surface generated species in the absence of collisions thus providing insight into surface reactions relevant to the deposition process. They observed two growth regimes: a pre-growth regime characterized by limited decomposition of the precursor and a growth regime in which near total consumption of the hexacarbonyl occurs. No evidence was found for recombinative desorption of C and O, providing independent support for conclusions drawn from the film studies. Summarizing, laser thermal deposition provides cleaner films than have previously been observed for cw and pulsed U V laser C V D of the hexacarbonyls. The temperatures obtained in the films are well below those predicted by surface scientists to be necessary for complete removal of C O . Under laser heating conditions the dissociation of C O is severely hindered at all but the lowest temperatures ( < 400 K ) .

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch021

2

6

6

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

21. SINGMASTER AND H O U L E

LCVD from the Metal Hexacarbonyls

301

ACKNOWLEDGMENTS The authors would like to thank Jorge Goitia for technical assistance. K A S gratefully acknowledges summer support from I B M East Fishkill, NSF (grant # RII-8911281), the donors of the Petroleum Research Fund (administered by the American Chemical Society) and the Affirmative Action Faculty Development Program at SJSU. This work was performed under a joint study agreement between I B M and SJSU.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ch021

REFERENCES 1) 2) 3) 4) 5) 6) 7) 8) 9) 10) 11) 12) 13) 14) 15)

16) 17) 18) 19)

20) 21) 22) 23) 24) 25)

Herman, I. P. Chem. Rev. 1989, 89, 1323. Solanki, R.; Boyer, P. K.; Collins, G. J. Appl. Phys. Lett. 1982, 41, 1048. Flynn, D. K.; Steinfeld, J. I.; Sethi, D. S. J. Appl. Phys., 1986, 59, 3914. Gluck, N. S.; Wolga, G. J.; Bartosch, C. E.; Ho, W.; Ying, Z. J. Appl. Phys., 1987, 61, 988. Jackson, R. L.; Tyndall, G. W. J. Appl. Phys., 1988, 64, 2092. Singmaster, Κ. Α.; Houle, F. Α.; Wilson, R. J. J. Phys. Chem., 1990, 89, 6864. Gilgen,H. H.; Cacouris, T.; Shaw, P. S.; Krchnavek, R. R.; Osgood, R. M. Appl. Phys., 1987, B42, 55. Turney, W.; James, S. G.; Cardinahl, P; grassian, V. H.; Singmaster, K. A. submitted Chem. of Mat. Petzold, H.C.; Putzar, R.; Weigmann, U.; Wilke, I. Mat. Res. Symp. Proc., 1988, 101, 75. Oprysko, M. M.; Beranek, M. W. J. Vac. Sci. Technol., 1987, B5, 496. Flitsch, F. Α.; Swanson, J. R.; Friend, C. M. Surf. Sci., 1991, 245, 85 and references therein. Singmaster, Κ. Α.; Houle, F. A. Mat. Res. Soc. Proc.,1991, 201, 159. Osgood, R. M.; Ehrlich, D. J. Opt. Lett., 1982, 7, 385. Singmaster, Κ. Α.; Houle, F. A. submitted to J. Phys. Chem. Creighton, J. R. Appl. Phys.,1986, 59, 410.; Gluck, N. S.; Ying, Z.; Bartosch, C. E.; Ho, W. J. Chem. Phys.,1987, 86, 4957.; Cho, C. C.; Bernasek, S. L.J. Vac. Sci. Technol., 1987, A5, 1088.; Germer, T. Α.; Ho, W. J. Chem. Phys.,1988, 89, 562. Ko, E. I.; Madix, R. J. Surf. Sci.,1981, 109, 221. Shinn, N. D.; Madey, T. E. J. Vac. Sci. Technol.,1985, A3, 1673. Kaplan, L. H.; d'Heurle, F. M. J. Electrochem. Soc., 1970, 117, 693. Yous, B.; Robin, S.; Robin, J.; Donnadieu, A. Thin Solid Films, 1985, 130, 181.; Carver, G. E.; Divrechy, Α.; Karbal, S.; Robin, J.; Donnadieu, A. Thin Solid Films, 1982, 94, 269. Bowker, M.; King, D. A. J. Chem.Soc.Faraday 1, 1980, 76, 758. Felder, T. E.; Estrup, P.J. Surf. Sci., 1978, 76, 464. Rigby, L. J. Can. J. Phys., 1964, 42, 1256. Benzinger, J. B.; Ko, E. I.; Madix, R. J. Catal., 1978, 54, 414. Umbach, E.; Menzel, D. Surf. Sci., 1983, 135, 199. Houle, F. Α.; Yeh, L. I. J. Phys. Chem. 1992, 96, 2691.

RECEIVED March 1, 1993

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ix001

Author Index Amrein, Andreas, 136 Armentrout, P. B., 61 Arnold, N., 250 Bàuerle, D., 250 Barnhart, Terence M., 116 Bartz, Jeffrey Α., 116 BelBruno, Joseph J., 49 Belyaev, Yu. E., 220 Blitz, Μ. Α., 177 Brown, Carl E., 96,177 Brum, Jeffrey L., 106 Chaiken, J., 1 Collings, Bruce Α., 136 Comita, Paul B., 279 Crim, F. Fleming, 116 Decker, S. Α., 177 DeLeon, Robert L., 188 Dem'yanenko, Α. V., 220 Deshmukh, Subhash, 106 Funk, Steve, 26 Galloway, Douglas B., 116 Garvey, James R, 188 Gerrity, Daniel P., 26 Glenewinkel-Meyer, Thomas, 1 Hackett, Peter Α., 96,136 Hossenlopp, Jeanne M., 75 Houle, F. Α., 292 House, P., 147

Huey, L. Gregory, 116 Ishikawa, Yo-ichi, 96 Jackson, Robert L., 122 Jacob, Wolfgang, 279 Kay, Eric, 279 Kimura, Katsumi, 86 Koplitz, Brent, 106 Kullmer, R., 250 McMahon, Robert J., 116 Mitchell,S. Α., 177 Mueller, Julie Α., 75 Peifer, William R., 188 Puretzky, Α. Α., 220 Rayner, David M., 96,136 Rice, Gary W., 273 Ryther, R. J., 147 Singmaster, Κ. Α., 292 Viegut, Terrance R., 75 Wang, Zhongrui, 106 Weiller, Bruce H., 164 Weisshaar, James C , 208 Weitz, Eric, 147 Wells, J. R., 147 Wight, Charles Α., 61 Xu, Xiaodong, 106 Yen, Yu-Fong, 106 Zhang, Rong, 122,279

Affiliation Index Aerospace Corporation, 164 Almaden Research Center, 122,279,292 Arvin/Calspan Corporation, 188 Dartmouth College, 49 Exxon R&E Company, 273 Institute for Molecular Science, 86 Johannes-Kepler-Universitàt Linz, 250 Marquette University, 75 Max-Planck-Institut fur Plasmaphysik, 279 National Research Council of Canada, 96,136,177

Northwestern University, 147 Reed College, 26 Russian Academy of Sciences, 220 San Jose State University, 292 State University of New York at Buffalo, 188 Syracuse University, 1 Tulane University, 106 University of Utah, 61 University of Wisconsin-Madison, 116,208

304

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

INDEX

305

Subject Index A

C

Agglomeration of metal particles, applications, 136 Alkylselenium compounds, multiphoton dissociation dynamics, 51-55 Arene-chromium tricarbonyls, parity-state selectivity in multiphoton dissociation, 26-47 Arene-Cr(CO) complexes, background for parity-state selectivity in multiphoton dissociation, 35-38/ Atoms, gas-phase formation, 220-247

Carbide synthesis using lasers, transition metal, description, 274-276 Chemical trapping experiments, 62 Chemical vapor deposition, laser-assisted, See Laser-assisted chemical vapor deposition Chemisorption on niobium clusters anticorrelation with ionization potential, 139-140 benzene reactions, 140 hydrogen chemisorption, 139 Classical photochemistry with lasers, 5 Clusters), gas-phase formation, 220-247 Cluster-assembled materials, descriptions, 136 Cluster synthesis using laser chemistry description, 12,15 history, 15 method for nascent cluster size distribution analysis, 16-21 size vs. properties, 15 CO elimination kinetic model of dynamics, 99-100 statistical models of dynamics, 98-99 CO vibrational- and rotational-state distributions, study, 62-63 C 0 laser, gas-phase formation of atoms, clusters, and ultrafine particle effect, 230,231/ Coordinatively unsaturated iron carbonyls, reactions, 151,153,154/, 156f Coordinatively unsaturated metal carbonyls bond dissociation energies, 160-161 bond dissociation energy measurement procedure, 153,155-159 experimental procedure, 149-151 future work, 162 iron carbonyl reactions, 151,153,154/,156r reaction kinetics for polynuclear iron carbonyls, 159-160 time-resolved spectra, 151,152/ Coordinatively unsaturated molecules, synthesis in laboratory, 189 Coordinatively unsaturated organometallic compounds importance of kinetic information, 148 kinetic measurement techniques, 148-149

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ix002

3

Β Binding, ligand, laser photoionization probes of effects in multiphoton dissociation of gas-phase transition metal complexes, 61-72 Bond dissociation energies coordinatively unsaturated metal carbonyls measurement procedure, 153,155-159 results, 160-161 niobium clusters, determination, 140-141 photodissociation of metal carbonyls, determination from branching ratios, 102,103/ Bond-selective chemistry, relevance to organometallics, 4 Branching ratios for photodissociation dynamics of metal carbonyls bond dissociation energy determination, 102,103/ fragment yields vs. photon energy, 101/ measurement, 102 Buffer gas, gas-phase formation of atoms, clusters, and ultrafine particle effect, 225,227-230 Buffer gas quenching as dynamic probe of organometallic photodissociation processes description, 123 experimental procedure, 123,125 photodissociation of dialkyl zincs, 124-130 photodissociation of ferrocene, 129,131-134

2

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

LASER CHEMISTRY OF ORGANOMETALLICS

306

Coordinatively unsaturated organotransition metal species, role in mechanistic organometallic chemistry, 188 Cr atom state distributions, study, 62 Cr(I) atomic states, 28-30/ emission spectra, 29,3\f,t excited state populations after multiphoton dissociation, 29,32-34 Cr(CO) multiphoton dissociation dynamics, 201-202 parity-state selectivity, 28-35 van der Waals clusters, photodissociation dynamics, 194-195

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ix002

6

D Delayed atomic luminescence, Rydberg mechanism, 230,232-235 Deposits, types for pyrolytic laser-induced chemical vapor deposition, 251-253/ Dialkyl zincs, photodissociation, 124-130 Di-n-butyl tin diacetate ^η,π*) chromophore excitation, 75-84 photolysis channels, 76,77/ use as esterification catalyst, 75 Diffusive regime, description, 17 Dimer formation in ultraviolet excitation ofMo(CO) luminescence pulse peak, 236-239 luminescence pulse tail, 237/239-240 process, 234 single-pulse excitation, 240-241 two-pulse experiments, 234-236 Direct dissociation processes, 3/4 Direct energy disposal measurements, photodissociation dynamics of metal carbonyls, 98 Direct writing for pyrolytic laser-induced chemical vapor deposition one-dimensional approach, 257-267 temperature distribution, 259-261 6

Ε

Electronic probes, description, 190 Electronic state(s), photoelectron determination, 87-90/ Electronic-state population, ligand effect, 89-91,93/ Electronic structure, theory, 209 Electronic structure of naked organometaUics in gas phase multiphoton dissociation dynamics of Cr(CO) , 201-204/ precursor temperature effect on 248-nm Cr(CO) multiphoton dissociation dynamics, 202,204/ probes, development, 200-201 time-resolved technique for molecular photoproduct detection, 202-203,205/ Energy-selective ionization, organometallic compounds in molecular beam, 116-120 6

6

F Fe(CO) -CO reaction, rate constant, 149 Fe(CO) van der Waals clusters, photodissociation dynamics, 191-192 Ferrocene, photodissociation, 129,131-134 Fluorescence techniques, detection and characterization of open-shell organometallics in gas phase, 188-205 Fluorocarbon polymer composite films, gold, laser-induced coalescence of gold clusters, 279-290 Free metal atoms and ions, production using gas-phase organometallic chemistry, 49-59 4

5

G Gas-phase formation of atoms, clusters, and ultrafine particles in ultraviolet excitation of metal carbonyls buffer gas effect, 225,227-230 buffer gas effect on M 0 luminescence spectrum, 243-247 C 0 laser effect, 230,231/ delayed atomic luminescence results, 223-226/ dimer formation from Mo(CO) , 234-241 experimental procedure, 220-222 Rydberg mechanism for delayed atomic luminescence, 230,232-235 2

2

6

Electron structure of naked organometallics in gas phase, implications for photodissociation dynamics, 203

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ix002

INDEX

307

Gas-phase formation of atoms, clusters, and ultrafine particles in ultraviolet excitation of metal carbonyls-Continued spectroscopic parameters, 222i,223 thermodynamic parameters, 222-223 ultrafine particle formation, 246-249 vibrational distribution in electronically excited state, 241,242/ Gas-phase kinetics of heterogeneous and homogeneous reactions models, 252,254/255 reaction flux effect, 255,256/ scanning electron micrograph, 255,257,258/ Gas-phase organometallic complexes, development of laser-based study techniques, 61-63 Gas-phase organometallic molecules, laser photoelectron spectroscopy, 86-94 Gas-phase photochemistry of organometallic molecules product detection methods, 116-117 studies, 116 Gas-phase transition metal atom(s), chemistry, 208 Gas-phase transition metal atom-small hydrocarbon reactions electronic structure theory, 208 experimental objective, 208-209 future work, 216 M reactions, 214-217/ M reactions, 210-212 M reactions, 212-213,217/ Gas-phase transition metal complexes, ligand binding effects in multiphoton dissociation, 61-72 Gas-to-particle conversion, development, 15 Gold clusters, laser-induced coalescence in gold fluorocarbon polymer composite films, 279-290 Group I-V semiconductor growth, site-specific laser chemistry, 106-114 Group V photochemistry, Η-atom production, 110-112/ Group VI metal hexacarbonyls advantages for film deposition by UV photolysis, 293 photochemical deposition, 294-299 thermal deposition, 296-300 +

H Η-atom production enhancement, 111,113/114 group V photochemistry, 110-112/ Haloethane photochemistry, site specific, description, 109-110 Heterogeneous metal carbonyl van der Waals clusters, photodissociation dynamics, 195-200 Hexacarbonyls, metal, laser-assisted chemical vapor deposition, 292-300 Homogeneity exponent, definition, 17 Homogeneous metal carbonyl van der Waals clusters, photodissociation dynamics, 191-192 Hydrocarbon-gas-phase transition metal atom reactions, small, See Gas-phase transition metal atom-small hydrocarbon reactions

Ionization potentials, niobium clusters, 138-139 Ionization techniques, detection and characterization of open-shell organometallics in gas phase, 188-205 Iron complexes, photodissociation, 87,88/ Iron molecules, multiphoton dissociation dynamics, 49-59

2 +

Κ Kernels, description, 16 Kinetic reaction, pyrolytic laser-induced chemical vapor deposition, 250-271 Kinetic studies, nickel atoms, 177-186

L Laser(s) advantages, applications, and functions, 273 information obtained about organometallic molecules, 116

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

July 21, 2012 | http://pubs.acs.org Publication Date: June 8, 1993 | doi: 10.1021/bk-1993-0530.ix002

308

LASER CHEMISTRY OF ORGANOMETALLICS

Laser-assisted chemical vapor deposition from metal hexacarbonyls experimental procedure, 293-294 photochemical deposition, 294-299 previous studies, 293 process, 292 reaction conditions, 292-293 thermal deposition, 296-300 Laser-based techniques, development for analysis of gas-phase organometallic complexes, 61 Laser chemistry of organometallics bond-selective chemistry, 4 cluster synthesis, 12,15-21 classical photochemistry, 5 commercial viability for synthesis, 4-5 correlation coefficients for ion signal vs. laser pulse energy, 12,14/ excitation sequence, energy-level representation, 7,9/ flow chart, 1,3/ history, 1,2 IR excitation, 2,4 isotope separation, 5 laser excitation, role, 2 ligand labilization, 4 multiphoton chemistry, 4 multiphoton dissociation of organoiron and organoselenium molecules, 49-59 multiple kinetic pathway, 5-6 one-color multiphoton excitation experiment, 7,8/ phases, 10 power index for formation of l ^ C r atoms, 7,11/12 power index for laser-induced excitation determination, 6-7 single-photon chemistry, 2,4 spatial position effect on power index, 12,13/ stages, 2 threshold for single-photon excitation, 6-7 Laser-driven synthesis of transition metal carbides, sulfides, and oxynitrides advantages and technique, 274 carbides, 274 oxynitrides, 276-277 sulfides, 276-277 Laser-induced chemical vapor deposition, 250

Laser-induced coalescence apparatus, 282 energy-dispersive X-ray analysis, 284,288/" film collapse vs. laser power, 284,286/ fluence, 284,289 maximum surface temperature, 288r,289 optical properties, 289-290 procedure, 282 resistivity of films, 284,285i scanning electron micrograph, 284,287/ time scale for complete collapse, 290 Laser-induced coalescence of gold clusters in goldfluorocarbonpolymer composite films advantages, 279 coalescence procedure, 282,284-290 film characterization, 282,283/285/ film preparation, 280-282 process, 279-281/ Laser-induced pyrolytic film decomposition of thin solid films, process, 279 Laser isotope separation, description, 5 Laser photodissociation of organometallic molecules, product detection methods, 116-117 Laser photoelectron spectroscopy of gas-phase organometallic molecules compact threshold photoelectron analyzer, 91-93/ high-resolution threshold photoelectron spectra, examples, 92,94/ ligand effect on electronic-state population, 89-91,93/ photodissociation of iron complexes, 87,88/ photoelectron determination of electronic states, 87-90/ Laser photoionization probes of ligand binding effects in multiphoton dissociation of gas-phase transition metal complexes branching between photodissociation and photoionization in VOCl , 70-72i nitric oxide ligand state distributions, 63-67 review objectives, 61 spin-state distributions of metal atoms, 66-71 Ligand(s), electronic-state population effect, 89-91,93/ Ligand binding, laser photoionization probes of effects in multiphoton dissociation of gas-phase transition metal complexes, 61-72 3

In Laser Chemistry of Organometallics; Chaiken, J.; ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

INDEX

309

Ligand-centered ^η,π*) chromophore excitation experimental procedure, 76-79 multiphoton excitation, 80-84 single-photon excitation, 79 single-photon fragmentation, 81-83 Ligand fragment state distributions, 62 Ligand labilization, role in organometallic photochemistry, 4 Liquid rare gas solvents, time-resolved IR spectroscopy of transient organometallics, 164-176

Metal-ligand chemistry, role of naked metal clusters, 137 Mo atoms, formation, 220 Mo buffer gas effect on luminescence spectrum, 243-247 spectroscopic and thermodynamic parameters, 222,223i Mo(CO) dimer formation in UV excitation, 234-241 spectroscopic and thermodynamic parameters, 222r vibrational distribution, 241,242/ Mo(CO) van der Waals clusters MS analysis of cluster beam, 192,193/ photodissociation dynamics, 192-194 possible origins of metal-oxide photoions, 192,194 Mo(CO) , spectroscopic and thermodynamic parameters, 222r Molecular beam studies, chemistry of niobium clusters, 138-141 Molecular photoproduct detection, timeresolved technique, 202-203,205/ Multiphoton association of gas-phase transition metal complexes, laser photoionization probes of ligand binding effect, 61-72 Multiphoton dissociation formation of metal atoms, 220 one-color experiment, 7,8/ power index for formation of P Cr atoms, 7,11/12 S